{"schema_version":"onlylabs.public_analysis_evidence.v1","title":"InclusionAI (Ant Group) analysis evidence pack","description":"Public onlylabs evidence pack for cited agent analysis: captured pages, ranked public signals, and stored web-search provenance used by the background analysis workflow.","url":"https://onlylabs.fyi/labs/inclusionai","json_url":"https://onlylabs.fyi/analysis/inclusionai/evidence.json","generated_at":"2026-06-11T18:08:28.073Z","org":{"slug":"inclusionai","name":"InclusionAI (Ant Group)","category":"neolab","category_label":"Neolab","dossier_url":"https://onlylabs.fyi/labs/inclusionai"},"analysis":null,"workflow":{"version":"onlylabs-deepagents-analysis-v3","provider":null,"model":null,"agent":null,"public_pack_mode":"local-pages-and-events","live_web_fetches":false,"note":"Public evidence exports do not trigger live Exa calls; stored Exa provenance is included when analysis metadata contains it."},"stats":{"pages":28,"events":118,"web":0,"evidence":88,"signal_desks":{"hiring":0,"forks":1,"releases":39,"talking":12,"repos":8},"data_radar_lanes":null,"data_radar_matches":null,"stored_analysis_evidence":null,"stored_analysis_web":null,"stored_analysis_signal_desks":null,"stored_analysis_data_radar_lanes":null,"stored_analysis_data_radar_matches":null},"stored_web_provenance":null,"evidence":[{"ref":"P1","kind":"page","title":"Ming-Omni-TTS: Simple and Efficient Unified Generation of Speech, Music, and Sound with Precise Control","date":"2026-06-11T03:23:10.711974+00:00","date_source":null,"source_url":"https://www.inclusion-ai.org/blog/ming-omni-tts","signal_url":null,"signal_json_url":null,"text":"Ming-Omni-TTS: Simple and Efficient Unified Generation of Speech, Music, and Sound with Precise Control | INCLUSION AI \n\nSkip to main content \nGITHUB 🤗 Hugging Face ｜ 🤖 ModelScope \n\nThe Introduction Video of Ming-Omni-TTS ​ \n\n🚀 Featured Abilities ​ \n\nMing-omni-tts is a high-performance unified audio generation model that achieves precise control over speech attributes and enables single-channel synthesis of speech, environmental sounds, and music. Powered by a custom 12.5Hz continuous tokenizer and Patch-by-Patch compression, it delivers competitive inference efficiency (3.1Hz). Additionally, the model features robust text normalization capabilities for the accurate and natural narration of complex mathematical and chemical expressions. \n\n🔊 Fine-grained Vocal Control: Enables precise control over speech rate, pitch, volume, emotion, and dialects via simple instructions. It achieves 93% accuracy for Cantonese and 46.7% for emotional control, outperforming CosyVoice3.\n\n🌌 Intelligent Voice Design: Features 100+ premium built-in voices and supports zero-shot voice design through natural language descriptions. Its performance on the Instruct-TTS-Eval-zh benchmark is on par with Qwen3-TTS.\n\n🎶 Immersive Unified Generation: The industry's first autoregressive model to jointly generate speech, ambient sound, and music in a single channel. Built on a custom 12.5Hz continuous tokenizer and a DiT head architecture, it delivers a seamless, \"in-the-scene\" auditory experience.\n\n⚡ High-efficiency Inference: Introduces a \"Patch-by-Patch\" compression strategy that reduces the LLM inference frame rate to 3.1Hz. This significantly cuts latency and enables podcast-style audio generation while preserving naturalness and audio detail.\n\n🧪 Professional Text Normalization: The model accurately parses and narrates complex formats, including mathematical expressions and chemical equations, ensuring natural-sounding output for specialized applications.\n\nModel Structure ​ \n\nMing-omni-tts is a unified audio language model for the generation of speech, music, and sound, based on a unified continuous audio tokenizer.\n\nUnified Continuous Audio Tokenizer. ​ \n\nUnified Audio Language Mo"},{"ref":"P2","kind":"page","title":"Ming-UniAudio: Speech LLM for Joint Understanding, Generation and Editing with Unified Representation","date":"2026-06-11T03:23:10.43785+00:00","date_source":null,"source_url":"https://www.inclusion-ai.org/blog/ming-uniaudio","signal_url":null,"signal_json_url":null,"text":"Ming-UniAudio: Speech LLM for Joint Understanding, Generation and Editing with Unified Representation | INCLUSION AI \n\nSkip to main content \nGITHUB 🤗 Hugging Face ｜ 🤖 ModelScope \n\nThe Introduction Video of Ming-UniAudio ​ \n\nAudio Edit Demo ​ \n\nEditing Tasks Video demos ​ \n\n🚀 Technical Highlights ​ \n\nFirst unified continuous speech tokenizer for both understanding and generation tasks: MingTok-Audio is a unified continuous speech tokenizer MingTok-Audio based on a VAE framework with a causal Transformer architecture, the first continuous speech tokenizer to effectively integrate semantic and acoustic features, and enables a closed-loop system with LLMs through hierarchical feature representations, makes it suitable for both understanding and generation tasks.\n\nFirst Speech LLM with unifed continuous tokenizer for both understanding and generation: Ming-UniAudio is an end-to-end unified speech language model with a single LLM backbone for both understanding and generation tasks, enhanced with a Diffusion Head to ensure high-fidelity speech synthesis.\n\nFirst universal free-form speech editing model for semantic and acoustic tasks without temporal regime: We introduce the first instruction-guided, free-form speech editing framework that supports comprehensive semantic and acoustic edits without requiring explicit edit regions, along with Ming-Freeform-Audio-Edit, the first open-source evaluation set for such tasks.\n\nFirst benchmark for free-form speech editing: We propose Audio-Edit-Benchmark, the first open-source free-form evaluation set comprising editing tasks of four semantic and five acoustic types, to evaluate the model's editing performance.\n\nInstruction-Guided Free-Form Speech Editing ​ \n\nSemantic Editing - Insert ​ \n\nInstruction Transcription Target Transcription Before Edit Speechedit Result \ninsert '简直' after the character or word at index 8. 真是个浪漫的邂逅可以说是英雄救美了 真是个浪漫的邂逅简直可以说是英雄救美了 \ninsert '真正' before the character or word '好'. 就有道而正焉可谓好学也已 就有道而正焉可谓真正好学也已 \ninsert 'clearly' before the character or word at index 8. Its legal status in Trinidad was insufficient to preserve its ecological status. Its legal status"},{"ref":"P3","kind":"page","title":"ABench: An Evolving Open-Source Benchmark","date":"2026-06-11T03:23:10.427618+00:00","date_source":null,"source_url":"https://www.inclusion-ai.org/blog/abench","signal_url":null,"signal_json_url":null,"text":"ABench: An Evolving Open-Source Benchmark | INCLUSION AI \n\nSkip to main content \nGITHUB \n🌟 Overview ​ \n\nABench is an evolving open-source benchmark suite designed to rigorously evaluate and enhance Large Language Models (LLMs) on complex cross-domain tasks . By targeting current model weaknesses, ABench provides systematic challenges in high-difficulty specialized domains , including physics, actuarial science, logical reasoning, law, and psychology.\n\n🎯 Core Objectives ​ \n\nAddress Evaluation Gaps : Design high-differentiation assessment tasks targeting underperforming question types \n\nEstablish Unified Standards : Create reliable, comparable benchmarks for multi-domain LLM evaluation\n\nExpand Capability Boundaries : Drive continuous optimization of knowledge systems and reasoning mechanisms through challenging innovative problems\n\n📊 Dataset Release Status ​ \n\nDomain Description Status \nPhysics 500 university/competition-level physics problems (400 static + 100 dynamic parametric variants) covering 10+ fields from classical mechanics to modern physics ✅ Released \nActuary Curated actuarial exam problems covering core topics: probability statistics, financial mathematics, life/non-life insurance, actuarial models, and risk management ✅ Released \nLogic High-differentiation logical reasoning problems from authoritative tests (LSAT/GMAT/GRE/SBI/Chinese Civil Service Exam) 🔄 In Preparation \nPsychology Psychological case studies and research questions (objective/subjective) evaluating understanding of human behavior and theories 🔄 In Preparation \nLaw Authoritative judicial exam materials covering core legal domains: criminal/civil/administrative/procedural/international law 🔄 In Preparation \n\n🌟 Overview \n🎯 Core Objectives \n📊 Dataset Release Status"},{"ref":"P4","kind":"page","title":"AWorld: The Agent Runtime for Self-Improvement","date":"2026-06-11T03:23:10.368719+00:00","date_source":null,"source_url":"https://www.inclusion-ai.org/blog/aworld","signal_url":null,"signal_json_url":null,"text":"AWorld: The Agent Runtime for Self-Improvement | INCLUSION AI \n\nSkip to main content \n\"Self-awareness: the hardest problem isn't solving within limits, it's discovering the own limitations\" \n\nTable of Contents ​ \n\nNews — Latest updates and announcements.\n\nIntroduction — Overview and purpose of the project.\n\nInstallation — Step-by-step setup instructions.\n\nQuick Start — Get started with usage examples.\n\nArchitecture — Explore the multi-agent system design.\n\nDemo — See the project in action with demonstrations.\n\nContributing — How to get involved and contribute.\n\nLicense — Project licensing details.\n\nNews ​ \n\n🦤 [2025/07/07] AWorld, as a runtime, is now ready for agentic training. See Self-Improvement section for details. We have updated our score to 77.08 on the GAIA test. Learn how to construct a GAIA runtime in the Demo section .\n\n🦩 [2025/06/19] We have updated our score to 72.43 on the GAIA test. Additionally, we have introduced a new local running mode. See ./README-local.md for detailed instructions.\n\n🐳 [2025/05/22] For quick GAIA evaluation, MCP tools, AWorld, and models are now available in a single Docker image. See ./README-docker.md for instructions and youtube video for demo.\n\n🥳 [2025/05/13] AWorld has updated its state management for browser use and enhanced the video processing MCP server, achieving a score of 77.58 on GAIA validation (Pass@1 = 61.8) and maintaining its position as the top-ranked open-source framework. Learn more: GAIA leaderboard \n\n✨ [2025/04/23] AWorld ranks 3rd on GAIA benchmark (69.7 avg) with impressive Pass@1 = 58.8, 1st among open-source frameworks. Reproduce with python examples/gaia/run.py \n\nIntroduction ​ \n\nAWorld (Agent World) is a multi-agent playground that enables agents to collaborate and self-improve. The framework supports a wide range of applications, including but not limited to product prototype verification, foundation model training and Multi-Agent System (MAS) design meta-learning.\n\nRuntime Key Features ​ \n\n1. Agent Construction 2. Topology Orchestration 3. Environments \n• ✅ Support for various model services \n• ✅ Integration with MCP tools \n• ✅ Custom tool support • ✅ Protocol encapsulation betwee"},{"ref":"P5","kind":"page","title":"inclusionAI/Ling repository metadata","date":"2026-06-11T03:18:02.664687+00:00","date_source":null,"source_url":"https://github.com/inclusionAI/Ling","signal_url":null,"signal_json_url":null,"text":"# inclusionAI/Ling\n\nDescription: Ling is a MoE LLM provided and open-sourced by InclusionAI. \n\nLanguage: Python\n\nLicense: MIT\n\nStars: 258\n\nForks: 25\n\nOpen issues: 2\n\nCreated: 2025-02-19T07:17:18Z\n\nPushed: 2025-05-14T06:34:57Z\n\nDefault branch: master\n\nFork: no\n\nArchived: no\n\nREADME:\n# Ling\n

\n\n

🤗 Hugging Face   |   🤖 ModelScope

\n\n## Introduction\n\nLing is a MoE LLM provided and open-sourced by InclusionAI. We introduce two different sizes, which are Ling-lite and Ling-plus. Ling-lite has 16.8 billion parameters with 2.75 billion activated parameters, while Ling-plus has 290 billion parameters with 28.8 billion activated parameters. Both models demonstrate impressive performance compared to existing models in the industry.\n\nTheir structure makes it easy to scale up and down and adapt to different tasks, so users can use these models for a wide range of tasks, from processing natural language to solving complex problems. Furthermore, the open-source nature of Ling promotes collaboration and innovation within the AI community, fostering a diverse range of use cases and enhancements.\n\nAs more developers and researchers engage with the platform, we can expect rapid advancements and improvements, leading to even more sophisticated applications. This collaborative approach accelerates development and ensures that the models remain at the forefront of technology, addressing emerging challenges in various fields.\n\n## Update\n\n- [2025-5-10] Ling-lite-1.5 has been released! It achieves significant progress in reasoning ability compared with previous Ling-lite. \n- [2025-4-15] Ling-lite is upgraded to Ling-lite-0415. The new model demonstrates notable improvements over its predecessor, Ling-lite-0220, especially on code and math.\n\n## Model Downloads\n\nYou can download the following table to see the various parameters for your use case. If you are located in mainland China, we also provide the model on ModelScope.cn to speed up the download process.\n\n| **Model** | *"},{"ref":"P6","kind":"page","title":"inclusionAI/PromptCoT repository metadata","date":"2026-06-11T03:18:02.486599+00:00","date_source":null,"source_url":"https://github.com/inclusionAI/PromptCoT","signal_url":null,"signal_json_url":null,"text":"# inclusionAI/PromptCoT\n\nDescription: A unified suite for generating elite reasoning problems and training high-performance LLMs, including pioneering attention-free architectures\n\nLanguage: Python\n\nLicense: MIT\n\nStars: 132\n\nForks: 15\n\nOpen issues: 4\n\nCreated: 2025-03-04T07:02:25Z\n\nPushed: 2026-01-31T06:55:27Z\n\nDefault branch: main\n\nFork: no\n\nArchived: no\n\nREADME:\n

PromptCoT 2.0

\n\n

\nScaling Prompt Synthesis for LLM Reasoning\n

\n\n

\n📄 Paper •\n🤗 Hugging Face\n

\n\n

\n\"PromptCoT\n

\n\n---\n\n## ✨ Overview\n\nPromptCoT 2.0 is a principled and scalable framework for **prompt synthesis** that substantially advances LLM reasoning in both **mathematics** and **programming**. \n\nIt introduces an **EM-style rationale-driven synthesis loop** (*concept → rationale → problem*), enabling the automatic generation of diverse and challenging problems at scale. These synthetic prompts support two complementary training regimes: \n\n**Self-Play**: the model improves autonomously by learning from verifiable signals (e.g., unit tests for code, boxed answers for math). With this approach, a **30B-A3B self-play model** achieves **92.1 on AIME24, 89.8 on AIME25, and 76.7 on HMMT Feb25**, as well as **74.2 on LiveCodeBench v5, 71.0 on v6, and 2079 Elo on Codeforces**. These results surpass strong open-source baselines (Qwen3-30B-A3B-Thinking) and achieve **competitive performance** with closed-source leaders such as Gemini 2.5 Pro and OpenAI o3 across math and code.\n\n**SFT**: a **7B** model trained **100% on synthetic data**—using prompts synthesized by PromptCoT 2.0 and **complete reasoning trajectories distilled from GPT-OSS-120B (medium)**—reaches **73.1 on AIME24, 65.6 on AIME25, and 1815 Elo on Codeforces**, outperforming counterparts trained on **human-written prompts**.\n\nUnleash the PromptCoT tide of reasoning!\n\n---\n\n## ⚡ Main Results\n\n**Self-Play @ Qwen3-30B-A3B-2507-Thinking:** \n\n

\n\n# AWorld: The Agent Harness for Your World\n\n\n\n

\n\n*\"The Next Frontier for AI is Your Expertise\"*\n\n[![Twitter Follow][twitter-image]][twitter-url]\n[![WeChat QR Code][wechat-image]][wechat-url]\n[![Discord][discord-image]][discord-url]\n[![License: MIT][license-image]][license-url]\n[![DeepWiki][deepwiki-image]][deepwiki-url]\n[![Tutorial][tutorial-image]][tutorial-url]\n\n\n\n

\n\n

\n\n[中文版](./README_zh.md) |\n[Automation](#your-journey-with-aworld-cli) |\n[Evolution](#evolution) |\n[Contributing](#contributing) |\n\n

\n\n---\n\n

\nGeneral AI often hits a \"wall of context\"—the nuanced data, workflows, and intuition that define your world. An agent's true power lies not in the model alone, but in its Agent Harness: the framework orchestrating its tools, memory, context, and execution.\n\nThis is the AWorld Thesis: A powerful harness is not enough. True AI scaling is unlocked only when experts like you embed the invaluable knowledge, effectively building the gate in that wall. \n\nAWorld is the platform designed for this singular purpose. We provide a complete, battle-tested Harness as the recipe for you, the expert, to forge your knowledge into a fleet of autonomous agents. Together, we move beyond AI's generic promise to create robust, precise applications that master your specific domain.\n

\n\n# From Expertise to Product\n\nSee what happens when expert knowledge is encoded into reusable **Skills**. The creations below are orchestrated by the AWorld Agent, demonstrating our core scaling law: as the community contributes more expertise, the entire ecosystem becomes more powerful.\n\nFrom **one-prompt video generation** to **deep-search workflows**, each example turns specialized know-how into r"},{"ref":"P8","kind":"page","title":"inclusionAI/.github repository metadata","date":"2026-06-11T03:18:02.103652+00:00","date_source":null,"source_url":"https://github.com/inclusionAI/.github","signal_url":null,"signal_json_url":null,"text":"# inclusionAI/.github\n\nStars: 0\n\nForks: 0\n\nOpen issues: 0\n\nCreated: 2025-03-24T08:35:16Z\n\nPushed: 2026-04-06T14:13:30Z\n\nDefault branch: main\n\nFork: no\n\nArchived: no\n\nREADME: none published or not readable through the GitHub API."},{"ref":"P9","kind":"page","title":"inclusionAI/Ring repository metadata","date":"2026-06-11T03:18:02.041445+00:00","date_source":null,"source_url":"https://github.com/inclusionAI/Ring","signal_url":null,"signal_json_url":null,"text":"# inclusionAI/Ring\n\nDescription: Ring is a reasoning MoE LLM provided and open-sourced by InclusionAI, derived from Ling. \n\nLanguage: Python\n\nLicense: MIT\n\nStars: 110\n\nForks: 2\n\nOpen issues: 2\n\nCreated: 2025-03-28T12:44:30Z\n\nPushed: 2025-08-05T05:48:56Z\n\nDefault branch: main\n\nFork: no\n\nArchived: no\n\nREADME:\n# Ring\n\n

\n\n

\n\n

\n🤗 Hugging Face   |   🤖 ModelScope\n\n## News\n* [2025-07]:🎉 Add [Ring-lite-2507](https://huggingface.co/inclusionAI/Ring-lite-2507) Model\n* [2025-06]:🎉 Add [Ring-lite](https://huggingface.co/inclusionAI/Ring-lite) Model\n* [2025-04]:🎉 Add [Ring-lite-linear-preview](hybrid_linear) Model\n\n## Introduction\n\nRing is a reasoning MoE LLM provided and open-sourced by InclusionAI, derived from [Ling](https://github.com/inclusionAI/Ling). We introduce Ring-lite-distill-preview, which has 16.8 billion parameters with 2.75 billion activated parameters. This model demonstrates impressive reasoning performance compared to existing models in the industry.\n\n## Model Downloads\n\nYou can download the following table to see the various parameters for your use case. If you are located in mainland China, we also provide the model on ModelScope.cn to speed up the download process.\n\n

\n\n| **Model** | **#Total Params** | **#Activated Params** | **Context Length** | **Download** |\n| :------------------: | :---------------: | :-------------------: | :----------------: | :--------------------------------------------------------------------------------------------------------------------------------------------------------: |\n| Ring-lite-2507 | 16.8B | 2.75B | 128K | [🤗 HuggingFace](https://huggingface.co/inclusionAI/Ring-lite-2507)
[🤖 ModelScope](https://modelscope.cn/models/inclusionAI/Ring-lite-2507) |\n| Ring-lite | 16.8B | 2.75B | 128K | [🤗 HuggingFace](https://huggingface.co/inclusionAI/Ring-lite)
[🤖 ModelScope](https://modelscope.cn/models/inclusionAI/Ring-lite) |\n| Ring-lite-distill-preview | 16.8B | 2.75B | 64K | [🤗 HuggingF"},{"ref":"P10","kind":"page","title":"inclusionAI/Ming repository metadata","date":"2026-06-11T03:18:01.818871+00:00","date_source":null,"source_url":"https://github.com/inclusionAI/Ming","signal_url":null,"signal_json_url":null,"text":"# inclusionAI/Ming\n\nDescription: Ming - facilitating advanced multimodal understanding and generation capabilities built upon the Ling LLM.\n\nLanguage: Jupyter Notebook\n\nLicense: MIT\n\nStars: 656\n\nForks: 58\n\nOpen issues: 25\n\nCreated: 2025-04-21T07:39:03Z\n\nPushed: 2026-03-17T11:55:16Z\n\nDefault branch: main\n\nFork: no\n\nArchived: no\n\nREADME:\n# Ming-flash-omni 2.0\n\n

\n\n

\n\n

📑 Technical Report｜🤗 Hugging Face｜ 🤖 ModelScope\n\n## Introduction\n\nThe newly released Ming-flash-omni 2.0 leverages the [Ling-2.0](https://github.com/inclusionAI/Ling-V2) architecture—a Mixture-of-Experts (MoE) framework comprising 100B total and 6B active parameters. Representing a generational advancement over its predecessor, it establishes new State-of-the-Art (SOTA) benchmarks among open-source omni-MLLMs. Ming-flash-omni 2.0 effectively synergizes foundational abilities with specialized domain expertise. In particular, it exhibits superior performance in visual encyclopedic knowledge, immersive speech synthesis, and high-dynamic image generation and manipulation.\n\n

\n\n

\n\n## 📌 Updates\n* [2026.02.11] 🔥 We release the official version of [Ming-flash-omni 2.0](https://mp.weixin.qq.com/s/hz2fsH1DGpp2zpY-Yngsog), an open-source SOTA omni-MLLM that pushes the boundaries of multimodal understanding and synthesis.\n* [2025.10.27] 🔥 We release the preview version of Ming-flash-omni：[Ming-flash-omni Preview](https://github.com/inclusionAI/Ming/tree/main).\n* [2025.07.15] 🔥 We release [Ming-lite-omni v1.5](https://github.com/inclusionAI/Ming/tree/v1.5) with significant improvements across all modalities.\n* [2025.06.12] 🔥 Our [Technical Report](https://arxiv.org/abs/2506.09344) is in public on arxiv.\n* [2025.05.28] 🔥 The off"},{"ref":"P11","kind":"page","title":"inclusionAI/ABench repository metadata","date":"2026-06-11T03:18:01.690674+00:00","date_source":null,"source_url":"https://github.com/inclusionAI/ABench","signal_url":null,"signal_json_url":null,"text":"# inclusionAI/ABench\n\nDescription: ABench is an evolving open-source benchmark suite designed to rigorously evaluate and enhance Large Language Models (LLMs) on complex cross-domain tasks.\n\nLanguage: Python\n\nLicense: Apache-2.0\n\nStars: 27\n\nForks: 1\n\nOpen issues: 1\n\nCreated: 2025-06-30T09:43:51Z\n\nPushed: 2026-04-17T11:04:50Z\n\nDefault branch: main\n\nFork: no\n\nArchived: no\n\nREADME:\n# ABench\n[![License: Apache 2.0](https://img.shields.io/badge/License-Apache%202.0-green.svg)](https://opensource.org/licenses/Apache-2.0)\n\n## 🌟 Overview\n\n**ABench** is an evolving open-source benchmark suite designed to rigorously evaluate and enhance Large Language Models (LLMs) on **complex cross-domain tasks**. By targeting current model weaknesses, ABench provides systematic challenges in **high-difficulty specialized domains**, including physics, actuarial science, logical reasoning, law, and psychology.\n\n## 🎯 Core Objectives\n1. **Address Evaluation Gaps**: Design high-differentiation assessment tasks targeting **underperforming question types**\n2. **Establish Unified Standards**: Create **reliable, comparable benchmarks** for multi-domain LLM evaluation\n3. **Expand Capability Boundaries**: Drive continuous optimization of knowledge systems and reasoning mechanisms through challenging innovative problems\n\n## 📊 Dataset Release Status\n\n| Domain | Description | Status |\n|----------------|------------------------------------------------------------------------------------------------------------------|------------------------------------|\n| **Physics** | 500 university/competition-level physics problems (400 static + 100 dynamic parametric variants) covering 10+ fields from classical mechanics to modern physics | [✅ Released](Physics/README.md) |\n| **Actuary** | Curated actuarial exam problems covering core topics: probability statistics, financial mathematics, life/non-life insurance, actuarial models, and risk management | [✅ Released](Actuary/README.md) |\n| **Logic** | High-differentiation logical reasoning problems from authoritative tests (LSAT/GMAT/GRE/SBI/Chinese Civil Service Exam) | [✅ Released](Logic/README.md) |\n| **Psychology** | Psychological case studies and research qu"},{"ref":"P12","kind":"page","title":"inclusionAI/AWorld-RL repository metadata","date":"2026-06-11T03:18:01.449392+00:00","date_source":null,"source_url":"https://github.com/inclusionAI/AWorld-RL","signal_url":null,"signal_json_url":null,"text":"# inclusionAI/AWorld-RL\n\nDescription: Agentic Learning Powered by AWorld\n\nLanguage: Python\n\nLicense: MIT\n\nStars: 110\n\nForks: 10\n\nOpen issues: 2\n\nCreated: 2025-07-01T07:52:11Z\n\nPushed: 2026-04-16T03:28:08Z\n\nDefault branch: main\n\nFork: no\n\nArchived: no\n\nREADME:\n

\n\n

\nAgentic Learning Powered by \"AWorldAWorld\n

\n\n
\n\n

\n arXiv(HardGen)\n arXiv(V2P) ｜\n arXiv(RAG-R1) ｜\n arXiv(FunReason) ｜\n arXiv(EnvTuning)｜\n arXiv(FunReason-MT)｜\n

\n\n

\n EnvTuning\n

\n\n## 📣 News\n\n[2026/04/06] 🎉🎉🎉[**FunReason (BalanceSFT)**](https://arxiv.org/html/2505.20192v3) was accepted as a finding paper of [ACL 2026](https://2026.aclweb.org/) conference!\n\n[2026/01/26] 🎉🎉🎉[**Environment Tuning**](https://arxiv.org/abs/2510.10197) was accepted at [ICLR 2026](https://iclr.cc/virtual/2026/poster/10007443) conference!\n\n[2026/01/04] 🔥🔥🔥[**HardGen**](./FunReason-MT) We propose **HadrGen**, an extension of the FunReason-MT.\n\n[2025/10/29] 🔥🔥🔥[**FunReason-MT**](./FunReason-MT) We propose **FunReason-MT**, a novel data synthesis framework designed to addres"},{"ref":"P13","kind":"page","title":"inclusionAI/M2-Reasoning repository metadata","date":"2026-06-11T03:18:01.410189+00:00","date_source":null,"source_url":"https://github.com/inclusionAI/M2-Reasoning","signal_url":null,"signal_json_url":null,"text":"# inclusionAI/M2-Reasoning\n\nDescription: M2-Reasoning: Empowering MLLMs with Unified General and Spatial Reasoning\n\nLanguage: Python\n\nStars: 48\n\nForks: 0\n\nOpen issues: 5\n\nCreated: 2025-07-02T12:48:46Z\n\nPushed: 2025-07-17T07:59:39Z\n\nDefault branch: main\n\nFork: no\n\nArchived: no\n\nREADME:\n# M2-Reasoning: Empowering MLLMs with Unified General and Spatial Reasoning\n\n📖 [Technical Report](./assets/M2-Reasoning.pdf) | 📄 [arXiv](https://arxiv.org/abs/2507.08306) | 🤗 [Hugging Face](https://huggingface.co/inclusionAI/M2-Reasoning)｜ 🤖 [ModelScope](https://www.modelscope.cn/models/inclusionAI/M2-Reasoning)\n\n## Introduction\n\nWe introduce M2-Reasoning-7B, a model designed to excel in both general and spatial reasoning. Our approach integrates two key innovations: (1) a novel data pipeline that generates 294.2K high-quality data samples (168K for cold-start fine-tuning and 126.2K for RLVR), which feature logically coherent reasoning trajectories and have undergone comprehensive assessment; and (2) a dynamic multi-task training strategy with step-wise optimization to mitigate conflicts between data, and task-specific rewards for delivering tailored incentive signals. This combination of curated data and advanced training allows M2-Reasoning-7B to set a new state-of-the-art (SOTA) across 8 benchmarks, showcasing superior performance in both general and spatial reasoning domains.\n![](assets/teaser.png)\n\n## 📌 Updates\n\n- [2025.07.14] 🔥 Our Technical Report is available on 📄 [arXiv](https://arxiv.org/abs/2507.08306).\n- [2025.07.11] 🔥 We release M2-Reasoning on 🤗 [Hugging Face](https://huggingface.co/inclusionAI/M2-Reasoning) and 🤖 [ModelScope](https://www.modelscope.cn/models/inclusionAI/M2-Reasoning).\n\n## Key Features\n\n- A High-quality Data Construction Pipeline: We design and implement a multi-stage data synthesis and curation pipeline that generates vast amounts of reasoning data.\n- A Dynamic Multi-Task Training Strategy: We propose a sophisticated training strategy that effectively handles data heterogeneity. It features step-wise dynamic optimization to mitigate conflicts between different data sources and a task-specific reward formulation to provide tailored incentive"},{"ref":"P14","kind":"page","title":"inclusionAI/inclusionAI.github.io repository metadata","date":"2026-06-11T03:18:01.264359+00:00","date_source":null,"source_url":"https://github.com/inclusionAI/inclusionAI.github.io","signal_url":null,"signal_json_url":null,"text":"# inclusionAI/inclusionAI.github.io\n\nDescription: 🔥 inclusionAI Official Website and Blog.\n\nLanguage: MDX\n\nLicense: MIT\n\nStars: 3\n\nForks: 6\n\nOpen issues: 0\n\nCreated: 2025-07-03T04:29:59Z\n\nPushed: 2026-05-20T04:06:57Z\n\nDefault branch: main\n\nFork: no\n\nArchived: no\n\nREADME:\n# inclusionAI Website\n\nOfficial website for [inclusionAI](https://github.com/inclusionAI), built with [Docusaurus 3](https://docusaurus.io/).\n\n## Prerequisites\n\n- Node.js ≥ 20\n- [pnpm](https://pnpm.io/) (used as the package manager)\n\n## Quick Start\n\n```bash\npnpm install\npnpm start\n```\n\n`pnpm start` launches a local dev server at `http://localhost:3000` with hot-reload.\n\n## Available Commands\n\n| Command | Description |\n| --------------------------- | --------------------------------------- |\n| `pnpm start` | Start local dev server (English locale) |\n| `pnpm start -- --locale zh` | Start dev server in Chinese locale |\n| `pnpm build` | Production build (output in `build/`) |\n| `pnpm serve` | Serve the production build locally |\n| `pnpm clear` | Clear Docusaurus cache |\n| `pnpm typecheck` | Run TypeScript type-checking |\n| `pnpm write-translations` | Extract i18n translation strings |\n\n## Adding a Blog Post\n\n### 1. Create the post directory and file\n\n```\nblog//index.mdx\n```\n\n### 2. Add required frontmatter\n\n```yaml\n---\ntitle: \"Your Post Title\"\ndate: 2025-01-01\nauthors: [inclusionai]\ntags: [Release, Insights]\n---\n```\n\n**Frontmatter fields:**\n\n| Field | Required | Description |\n| ----------------- | -------- | ------------------------------------------------- |\n| `title` | Yes | Post title (shown in listings and page ``) |\n| `date` | Yes | Publication date (`YYYY-MM-DD`) |\n| `authors` | Yes | Author key(s) from `blog/authors.yml` |\n| `tags` | No | Tag list — certain tags control which site sections the post appears in (see below) |\n| `draft` | No | Set `true` to hide from build output |\n| `custom_edit_url` | No | Set `null` to hide the \"Edit this page\" link |\n\n### Tag Reference\n\nTags serve two purposes: they render as colored badges in the Blog listing and they route posts into dedicated sections on the [Research page](/research).\n\n**Special routing tags** (case-insensitive):\n\n| Tag"},{"ref":"P15","kind":"page","title":"inclusionAI/ASearcher repository metadata","date":"2026-06-11T03:18:01.005182+00:00","date_source":null,"source_url":"https://github.com/inclusionAI/ASearcher","signal_url":null,"signal_json_url":null,"text":"# inclusionAI/ASearcher\n\nDescription: An Open-Source Large-Scale Reinforcement Learning Project for Search Agents\n\nLanguage: Python\n\nStars: 594\n\nForks: 38\n\nOpen issues: 14\n\nCreated: 2025-08-05T07:42:30Z\n\nPushed: 2025-11-26T06:28:16Z\n\nDefault branch: main\n\nFork: no\n\nArchived: no\n\nREADME:\n<h1 align=\"center\">\n<em>ASearcher</em>: An Open-Source Large-Scale\nReinforcement Learning Project for Search Agents\n</h1>\n\n<p align=\"center\">| <a href=\"https://arxiv.org/abs/2508.07976\"><b>📰 Paper</b></a> | <a href=\"https://huggingface.co/datasets/inclusionAI/ASearcher-train-data\"><b>🤗 Datasets</b></a> | <a href=\"https://huggingface.co/collections/inclusionAI/asearcher-6891d8acad5ebc3a1e1fb2d1\"><b>🤗 Models</b></a> | </p>\n\n# Introduction\n\nASearcher is an open-source framework designed for large-scale online reinforcement learning (RL) training of search agents. Our mission is to advance Search Intelligence to expert-level performance. We are fully committed to open-source by releasing model weights, detailed training methodologies, and data synthesis pipelines. Additionally, we provide comprehensive guidance on building and training customized agents based on AReaL. ASearcher empowers developers to build their own high-performance search agents easily and cost-effectively.\n\n**ASearcher Highlights**\n\n+ 🔁 **Data Synthesis Agent**: We introduce a prompt-based LLM agent that autonomously generates grounded, challenging, and highly uncertain QA pairs to enhance training diversity.\n+ ⚡ **Fully Asynchronous Agentic RL**: Our scalable agentic RL framework decouples trajectory collection from model training, eliminating GPU idle time and enabling efficient long-horizon RL training.\n+ 🌐 **RL Enables Long-Horizon Search**: Through RL training, ASearcher exhibits long-horizon search, with tool calls exceeding 100 rounds and generated tokens surpassing 400k during RL training. \n+ 🏆 **Cutting-Edge Performance**: With a simple agent design and no external LLMs, ASearcher achieves *Avg@4 scores of 58.7, 51.1, and 74.5* on GAIA, xBench-DeepSearch, and Frames, respectively, surpassing other open-source search agents on the same 32B scale. ASearcher achieves *Pass@4 scores of 74.7, 75.0, and 8"},{"ref":"P16","kind":"page","title":"inclusionAI/GroveMoE repository metadata","date":"2026-06-11T03:18:00.906544+00:00","date_source":null,"source_url":"https://github.com/inclusionAI/GroveMoE","signal_url":null,"signal_json_url":null,"text":"# inclusionAI/GroveMoE\n\nLanguage: Python\n\nStars: 25\n\nForks: 1\n\nOpen issues: 1\n\nCreated: 2025-08-12T02:48:59Z\n\nPushed: 2025-08-20T03:29:33Z\n\nDefault branch: master\n\nFork: no\n\nArchived: no\n\nREADME:\n<div align=\"center\">\n<h1><strong>GroveMoE</strong></h1>\n</div>\n<!-- [![arXiv](https://img.shields.io/badge/arXiv-2508.07785-b31b1b.svg)](https://arxiv.org/abs/2508.07785)\n[![Hugging Face](https://img.shields.io/badge/Hugging%20Face-FFD21E?logo=huggingface&logoColor=000)](https://huggingface.co/inclusionAI) -->\n\n<p align=\"center\">\n🤗 <a href=\"https://huggingface.co/collections/inclusionAI/grovemoe-68a2b58acbb55827244ef664\">Models</a>   |    📑 <a href=\"https://arxiv.org/abs/2508.07785\">Paper</a>    |    🔗 <a href=\"https://github.com/inclusionAI/GroveMoE\">Github</a>  \n\n## Overview\n\nGroveMoE is an **open-source** family of large language models developed by the **AGI Center, Ant Group Research** that introduces **Grove MoE**, a new sparse architecture using **adjugate experts** for dynamic computation allocation. \nWith **33 B total parameters** and **3.14–3.28 B active parameters per token**, GroveMoE delivers **state-of-the-art** results across reasoning, mathematics, and code generation while keeping inference costs low. \n\n<p align=\"center\"><img src=\"assets/grovemoe.png\" width=\"95%\"></p>\n\n---\n\n## Key Highlights\n| Feature | Description |\n|---------|-------------|\n| **Architecture** | Novel **adjugate experts** grouped with ordinary experts; shared computation is executed once, then reused, cutting FLOPs. |\n| **Sparse Activation** | 33 B params total, only **3.14–3.28 B active** per token. |\n| **Training** | Mid-training + SFT, up-cycled from **Qwen3-30B-A3B-Base**; preserves prior knowledge while adding new capabilities. |\n| **Open** | Weights, configs will be fully released under Apache 2.0 upon approval. |\n\n---\n\n## Run GroveMoE\n\n### 🤗 Transformers Quick Start\nTransformers is a library of pretrained natural language processing for inference and training. \n\nThe following contains a code snippet illustrating how to use GroveMoE to generate content based on given inputs. \n```python\nfrom transformers import AutoModelForCausalLM, Auto"},{"ref":"P17","kind":"page","title":"inclusionAI/MoBE repository metadata","date":"2026-06-11T03:18:00.687426+00:00","date_source":null,"source_url":"https://github.com/inclusionAI/MoBE","signal_url":null,"signal_json_url":null,"text":"# inclusionAI/MoBE\n\nDescription: Mixture-of-Basis-Experts for Compressing MoE-based LLMs\n\nLanguage: Python\n\nStars: 34\n\nForks: 5\n\nOpen issues: 2\n\nCreated: 2025-08-13T08:19:35Z\n\nPushed: 2025-12-24T15:41:10Z\n\nDefault branch: main\n\nFork: no\n\nArchived: no\n\nREADME:\n<h1 align=\"center\">\nMoBE: Mixture-of-Basis-Experts for Compressing <br> MoE-based LLMs\n</h1>\n<div style=\"text-align:center;\">\n<a href=\"https://arxiv.org/abs/2508.05257\">\n<img src=\"https://img.shields.io/badge/arXiv-2508.05257-b31b1b.svg\" alt=\"arXiv\">\n</a>\n</div>\n\n---\n## ✅ Feature List\n- [x] **Supported multiple MoE models (BF16)**\n- [x] Ling Family\n- [x] Qwen3MoE Family\n- [x] DeepSeek-V3\n- [x] Kimi-K2-Instruct\n- [ ] **Supported SGLang inference (with fused-MoE kernel)**\n- [ ] **Supported MoBE mega-kernel (high-performance fused kernel for MoBE)**\n> 💡 *Coming soon: Optimized inference kernels for MoBE models to maximize throughput and memory efficiency.*\n---\n\n## 📘 Introduction\n\n**MoBE (Mixture-of-Basis-Experts)** is a novel model compression technique designed for MoE LLMs developed by the **AGI Center, Ant Group Research**. It achieves efficient parameter reduction by factorizing each expert's weight matrix as:\n\n$$\n\\mathbf{W} = \\mathbf{A}\\mathbf{B}, \\quad \\text{where} \\quad \\mathbf{B} = \\sum_{i=1}^m \\alpha_i B_i\n$$\n\n- $\\mathbf{A}$: Expert-specific matrix \n- $\\mathbf{B}$: Linear combination of **basis matrices** across all experts, weighted by coefficients $\\alpha_i$\n\nThe factorization is learned by minimizing the **reconstruction error** between the original and compressed weight matrices.\n\n### 🔍 Key Results\nMoBE significantly outperforms prior compression methods with minimal accuracy degradation:\n- Reduces parameter count by **24%–30%** in leading open-source models\n- Incurs only **1%–2% absolute accuracy drop** (≈2% relative)\n- Demonstrated on **Qwen3-235B**, **DeepSeek-V3 (671B)**, and **Kimi-K2-Instruct (1T)**\n\n## 📊 Evaluation Results\n\n![results](results.jpg)\n---\n\n## 🚀 Quickstart\n\n### 🔧 Installation\n```\npip install -r requirements.txt\n```\n\n---\n\n### 🛠️ Step-by-Step Instructions\nConverting an MoE model to MoBE involves two stages:\n1. **Train** the MoBE decomposition.\n2. **Generate** either a nativ"},{"ref":"P18","kind":"page","title":"inclusionAI/UI-Venus repository metadata","date":"2026-06-11T03:18:00.613795+00:00","date_source":null,"source_url":"https://github.com/inclusionAI/UI-Venus","signal_url":null,"signal_json_url":null,"text":"# inclusionAI/UI-Venus\n\nDescription: UI-Venus is a native UI agent designed to perform precise GUI element grounding and effective navigation using only screenshots as input.\n\nLanguage: Python\n\nStars: 1010\n\nForks: 85\n\nOpen issues: 12\n\nCreated: 2025-08-15T11:57:47Z\n\nPushed: 2026-05-11T06:54:30Z\n\nDefault branch: UI-Venus-1.5\n\nFork: no\n\nArchived: no\n\nREADME:\n<h1 align=\"center\">\n<img src=\"assets/ui-venus-logo-3.png\" width=\"60\" align=\"center\"> UI-Venus 1.5\n</h1>\n\n<p align=\"center\">\n<a href=\"https://opensource.org/licenses/Apache-2.0\"><img src=\"https://img.shields.io/badge/License-Apache_2.0-blue.svg\" alt=\"License\"></a>\n<a href=\"https://arxiv.org/abs/2602.09082\"><img src=\"https://img.shields.io/badge/Report-Technical%20Report-blueviolet?logo=notion\" alt=\"Report\"></a>\n<a href=\"https://ui-venus.github.io/UI-Venus-1.5/\"><img src=\"https://img.shields.io/badge/🌐%20Website-UI--Venus--1.5-blue\" alt=\"Website\"></a>\n<a href=\"https://github.com/inclusionAI/UI-Venus\"><img src=\"https://img.shields.io/badge/GitHub-Repository-green?logo=github\" alt=\"GitHub\"></a>\n<a href=\"https://huggingface.co/collections/inclusionAI/ui-venus-689f2fb01a4234cbce91c56a\"><img src=\"https://img.shields.io/badge/Hugging%20Face-Model-orange?logo=huggingface\" alt=\"Hugging Face\"></a>\n</p>\n\n<p align=\"center\">\n<em>UI-Venus 1.5 is a unified, end-to-end GUI Agent designed for robust real-world applications. The model family includes two dense (2B/8B) and one MoE (30B-A3B) variants to meet various downstream scenarios.</em>\n</p>\n\n**Upgrades from UI-Venus 1.0:**\n- 🔹 **Mid-Training Stage**: 10B tokens across 30+ datasets for foundational GUI semantics\n- 🔹 **Online RL**: Full-trajectory rollouts for long-horizon dynamic navigation\n- 🔹 **Model Merging**: Unified agent combining grounding, web, and mobile specialists\n\n**Results:** SOTA on ScreenSpot-Pro (69.6%), VenusBench-GD (75.0%), AndroidWorld (77.6%), with robust navigation across 40+ Chinese mobile apps.\n\n---\n\n<p align=\"center\">\n📈 <strong>UI-Venus Benchmark Performance</strong>\n</p>\n\n<p align=\"center\">\n<img src=\"assets/performance_venus.png\" alt=\"UI-Venus Performance\" width=\"1200\" />\n</p>\n\n> **Figure:** Performance of UI-Venus 1.5 across multiple benchmarks"},{"ref":"P19","kind":"page","title":"inclusionAI/Ling-V2 repository metadata","date":"2026-06-11T03:18:00.248367+00:00","date_source":null,"source_url":"https://github.com/inclusionAI/Ling-V2","signal_url":null,"signal_json_url":null,"text":"# inclusionAI/Ling-V2\n\nDescription: Ling-V2 is a MoE LLM provided and open-sourced by InclusionAI.\n\nLanguage: Python\n\nLicense: MIT\n\nStars: 271\n\nForks: 17\n\nOpen issues: 6\n\nCreated: 2025-09-04T06:54:36Z\n\nPushed: 2025-10-04T06:15:38Z\n\nDefault branch: main\n\nFork: no\n\nArchived: no\n\nREADME:\n# Ling-V2\n<p align=\"center\"><img src=\"./figures/ant-bailing.png\" width=\"100\"/></p>\n\n<p align=\"center\">🤗 <a href=\"https://huggingface.co/inclusionAI\">Hugging Face</a>   |   🤖 <a href=\"https://modelscope.cn/organization/inclusionAI\">ModelScope</a></p>\n\n## Introduction\n\nToday, we are excited to announce the open-sourcing of __Ling 2.0__ — a family of MoE-based large language models that combine __SOTA performance__ with __high efficiency__.\nThe first released version, Ling-mini-2.0, is compact yet powerful. It has __16B total parameters__, but only __1.4B__ are activated per input token (non-embedding 789M). Trained on more than __20T tokens__ of high-quality data and enhanced through multi-stage supervised fine-tuning and reinforcement learning, Ling-mini-2.0 achieves remarkable improvements in complex reasoning and instruction following. With just 1.4B activated parameters, it still reaches the top-tier level of sub-10B dense LLMs and even matches or surpasses much larger MoE models.\n\n<p align=\"center\"><img src=\"./figures/ling-miniv2-eval.png\" /></p>\n\n### Strong General and Professional Reasoning\n\nWe evaluated Ling-mini-2.0 on challenging general reasoning tasks in coding (LiveCodeBench, CodeForces) and mathematics (AIME 2025, HMMT 2025), as well as knowledge-intensive reasoning tasks across multiple domains (MMLU-Pro, Humanity's Last Exam). Compared with sub-10B dense models (e.g., Qwen3-4B-instruct-2507, Qwen3-8B-nothinking) and larger-scale MoE models (Ernie-4.5-21B-A3B-PT, GPT-OSS-20B/low), Ling-mini-2.0 demonstrated outstanding overall reasoning capabilities.\n\n### 7× Equivalent Dense Performance Leverage\n\nGuided by Ling Scaling Laws, Ling 2.0 adopts a __1/32 activation ratio__ MoE architecture, with empirically optimized design choices in expert granularity, shared expert ratio, attention ratio, aux-loss free + sigmoid routing strategy, MTP loss, QK-Norm, half"},{"ref":"P20","kind":"page","title":"inclusionAI/Ring-V2 repository metadata","date":"2026-06-11T03:18:00.180783+00:00","date_source":null,"source_url":"https://github.com/inclusionAI/Ring-V2","signal_url":null,"signal_json_url":null,"text":"# inclusionAI/Ring-V2\n\nDescription: Ring-V2 is a reasoning MoE LLM provided and open-sourced by InclusionAI.\n\nLanguage: Python\n\nLicense: MIT\n\nStars: 99\n\nForks: 8\n\nOpen issues: 1\n\nCreated: 2025-09-22T15:49:26Z\n\nPushed: 2025-10-23T06:26:07Z\n\nDefault branch: main\n\nFork: no\n\nArchived: no\n\nREADME:\n# Ring-V2\n<p align=\"center\"><img src=\"./figures/ant-bailing.png\" width=\"100\"/></p>\n\n<p align=\"center\">🤗 <a href=\"https://huggingface.co/inclusionAI\">Hugging Face</a>   |   🤖 <a href=\"https://modelscope.cn/organization/inclusionAI\">ModelScope</a></p>\n\n## News\n* [2025-10]:🎉 Add [Ring-linear-2.0](https://arxiv.org/abs/2510.19338) 📖 Technical Report\n* [2025-10]:🎉 Add [Ring-1T](https://huggingface.co/inclusionAI/Ring-1T) Model\n* [2025-09]:🎉 Add [Ring-linear-2.0](https://github.com/inclusionAI/Ring-V2/tree/main/hybrid_linear) Series\n* [2025-09]:🎉 Add [Ring-flash-2.0](https://huggingface.co/inclusionAI/Ring-flash-2.0) Model\n* [2025-09]:🎉 Add [Ring-mini-2.0](https://huggingface.co/inclusionAI/Ring-mini-2.0) Model\n\n## Introduction\nRing-V2 is a family of reasoning MoE LLMs with a range of sizes provided and open-sourced by InclusionAI, derived from [Ling-V2](https://github.com/inclusionAI/Ling-V2). These models achieve leading performance in complex reasoning at similar sizes, while maintaining high inference speed thanks to their highly sparse architecture.\n\n## Model Downloads\n\n| **Model** | **Context Length** | **Download** |\n|:----------------------:| :----------------: |:--------------------------------------------------------------------------------------------------------------------------------------------------------------------:|\n| Ring-1T | 64K -> 128K (YaRN) | [🤗 HuggingFace](https://huggingface.co/inclusionAI/Ring-1T) <br>[🤖 ModelScope](https://www.modelscope.cn/models/inclusionAI/Ring-1T) |\n| Ring-1T-FP8 | 64K -> 128K (YaRN) | [🤗 HuggingFace](https://huggingface.co/inclusionAI/Ring-1T-FP8) <br>[🤖 ModelScope](https://www.modelscope.cn/models/inclusionAI/Ring-1T-FP8) |\n| Ring-flash-2.0 | 32K -> 128K (YaRN) | [🤗 HuggingFace](https://huggingface.co/inclusionAI/Ring-flash-2.0) <br>[🤖 ModelScope](https://www.modelscope.cn/models/inclusionAI/Ring-f"},{"ref":"P21","kind":"page","title":"inclusionAI/Ming-Freeform-Audio-Edit repository metadata","date":"2026-06-11T03:17:59.935674+00:00","date_source":null,"source_url":"https://github.com/inclusionAI/Ming-Freeform-Audio-Edit","signal_url":null,"signal_json_url":null,"text":"# inclusionAI/Ming-Freeform-Audio-Edit\n\nLanguage: Python\n\nStars: 16\n\nForks: 2\n\nOpen issues: 0\n\nCreated: 2025-09-29T03:19:43Z\n\nPushed: 2025-10-27T07:46:37Z\n\nDefault branch: main\n\nFork: no\n\nArchived: no\n\nREADME:\n# README\n\n## Introduction\nThis repository hosts Ming-Freeform-Audio-Edit, the benchmark test set for evaluating the downstream editing tasks of the Ming-UniAudio model.\n\nThis test set covers 7 distinct editing tasks, categorized as follows:\n\n+ Semantic Editing (3 tasks):\n\n+ Free-form Deletion\n+ Free-form Insertion\n+ Free-form Substitution\n+ Acoustic Editing (5 tasks):\n+ Time-stretching\n+ Pitch Shifting\n+ Dialect Conversion\n+ Emotion Conversion\n+ Volume Conversion\n\nThe audio samples are sourced from well-known open-source datasets, including seed-tts eval, LibriTTS, and Gigaspeech.\n\n## Dataset statistics\n### Semantic Editing\n#### full version\n| Task Types\\ # samples \\ Language | Zh deletion | Zh insertion | Zh substitution | En deletion | En insertion | En substitution |\n| -------------------------------- | ----------: | -----------: | --------------: | ----------: | -----------: | --------------: |\n| Index-based | 186 | 180 | 36 | 138 | 100 | 67 |\n| Content-based | 95 | 110 | 289 | 62 | 99 | 189 |\n| Total | 281 | 290 | 325 | 200 | 199 | 256 |\n\n#### basic version\n| Task Types\\ # samples \\ Language | Zh deletion | Zh insertion | Zh substitution | En deletion | En insertion | En substitution |\n| -------------------------------- | ----------: | -----------: | --------------: | ----------: | -----------: | --------------: |\n| Index-based | 92 | 65 | 29 | 47 | 79 | 29 |\n| Content-based | 78 | 105 | 130 | 133 | 81 | 150 |\n| Total | 170 | 170 | 159 | 180 | 160 | 179 |\n\n*Index-based* instruction: specifies an operation on content at positions *i* to *j*. (e.g. delete the characters or words from index 3 to 12)\n\n*Content-based*: targets specific characters or words for editing. (e.g. insert 'hello' before 'world')\n### Acoustic Editing\n| Task Types\\ # samples \\ Language | Zh | En |\n| -------------------------------- | ---: | ---: |\n| Time-stretching | 50 | 50 |\n| Pitch Shifting | 50 | 50 |\n| Dialect Conversion | 250 | --- |\n| Emotion Conversion | 84 | 72 |\n| Volume C"},{"ref":"P22","kind":"page","title":"inclusionAI/MingTok-Audio repository metadata","date":"2026-06-11T03:17:59.8319+00:00","date_source":null,"source_url":"https://github.com/inclusionAI/MingTok-Audio","signal_url":null,"signal_json_url":null,"text":"# inclusionAI/MingTok-Audio\n\nLanguage: Python\n\nLicense: MIT\n\nStars: 88\n\nForks: 9\n\nOpen issues: 4\n\nCreated: 2025-09-29T03:19:13Z\n\nPushed: 2026-02-24T04:10:19Z\n\nDefault branch: main\n\nFork: no\n\nArchived: no\n\nREADME:\n<p align=\"center\">\n<img src=\"./assets/ant-bailing.png\" width=\"100\"/>\n<p>\n\n<p align=\"center\">📝<a href=\"https://arxiv.org/abs/2511.05516\">Technical Report</a> 📖<a href=\"https://xqacmer.github.io/Ming-Unitok-Audio.github.io\">Project Page</a> ｜🤗 <a href=\"https://huggingface.co/inclusionAI/MingTok-Audio\">Hugging Face</a>｜ 🤖 <a href=\"https://modelscope.cn/models/inclusionAI/MingTok-Audio\">ModelScope</a>\n\n## Architecture\n<!-- ![MingTok-Audio](assets/uniaudio-tokenizer.png)\n![MingTok-Audio-training](assets/uniaudio-tokenizer-training.png) -->\n\n<p align=\"center\">\n<img src=\"assets/uniaudio-tokenizer.png\" alt=\"MingTok-Audio\"/>\n</p>\n\n## Key Features\n- 🚀 **First Unified Continuous Speech Tokenizer:** the first continuous audio tokenizer to effectively integrate semantic and acoustic features, suitable for both understanding and generation tasks.\n- 🎧 **High-Quality Reconstruction:** Achieve high-quality audio generation by modeling continuous features with a VAE, minimizing information loss and preserving intricate acoustic textures.\n- 🌐 **Convolution-Free Efficiency:** Built on a pure causal transformer architecture, completely eliminating convolutional layers for superior efficiency and a simpler design.\n\n## Installation\n```\npip install -r requirements.txt\n```\n\n## Quick start\n```python\nimport torch\nimport torchaudio\n\nfrom audio_tokenizer.modeling_audio_vae import AudioVAE\n\nmodel = AudioVAE.from_pretrained('inclusionAI/MingTok-Audio')\nmodel = model.cuda()\nmodel.eval()\n\nwaveform, sr = torchaudio.load('data/1089-134686-0000.flac', backend='soundfile')\nsample = {'waveform': waveform.cuda(), 'waveform_length': torch.tensor([waveform.size(-1)]).cuda()}\n\nwith torch.no_grad():\nwith torch.autocast(device_type='cuda', dtype=torch.bfloat16):\nlatent, frame_num = model.encode_latent(**sample)\noutput_waveform = model.decode(latent)\n\ntorchaudio.save('./1089-134686-0000_reconstruct.wav', output_waveform.cpu()[0], sample_rate=16000)\n```\n\n## Performance\n### Speech reconstruct"},{"ref":"P23","kind":"page","title":"inclusionAI/Ming-UniAudio repository metadata","date":"2026-06-11T03:17:59.729192+00:00","date_source":null,"source_url":"https://github.com/inclusionAI/Ming-UniAudio","signal_url":null,"signal_json_url":null,"text":"# inclusionAI/Ming-UniAudio\n\nDescription: Ming-UniAudio: Speech LLM for Joint Understanding, Generation and Editing with Unified Representation\n\nLanguage: Python\n\nLicense: MIT\n\nStars: 448\n\nForks: 30\n\nOpen issues: 8\n\nCreated: 2025-09-29T03:23:18Z\n\nPushed: 2025-11-27T02:51:18Z\n\nDefault branch: main\n\nFork: no\n\nArchived: no\n\nREADME:\n# Ming-UniAudio\n\n<p align=\"center\">\n<img src=\"./figures/ant-bailing.png\" width=\"100\"/>\n<p>\n\n<p align=\"center\">📝<a href=\"https://arxiv.org/abs/2511.05516\">Technical Report</a> ｜🌐<a href=\"https://xqacmer.github.io/Ming-Unitok-Audio.github.io/\">Project Page</a> ｜🤗 <a href=\"https://huggingface.co/inclusionAI/Ming-UniAudio-16B-A3B\">Hugging Face</a>｜ 🤖 <a href=\"https://modelscope.cn/models/inclusionAI/Ming-UniAudio-16B-A3B\">ModelScope</a>\n\n## Table of Contents\n- [Introduction](#introduction)\n- [Updates](#updates)\n- [Key Features](#key-features)\n- [Evaluation](#evaluation)\n- [Speech Tokenizer](#speech-tokenizer)\n- [Speech Understanding](#speech-understanding)\n- [Speech Generation](#speech-generation)\n- [Speech Editing](#speech-editing)\n- [Model & Benchmark Downloads](#model--benchmark-downloads)\n- [Environment Preparation](#environment-preparation)\n- [Example Usage](#example-usage)\n- [SFT](#sft)\n- [Citation](#citation)\n- [Join Us](#join-us)\n\n## Introduction\n\nMing-UniAudio is a novel framework that unifies speech understanding, generation, and editing. Its core is a unified continuous speech tokenizer that effectively unifies semantic and acoustic features within an end-to-end model. We developed a speech language model that strikes a balance between generation and understanding capabilities based on the unified continuous audio tokenizer. Leveraging this foundational model, which exhibits robust performance in both domains, we further trained a dedicated speech editing model built upon [Ming-Lite-Omni](https://github.com/inclusionAI/Ming). Crucially, Ming-UniAudio is the first to enable universal, free-form speech editing guided solely by natural language instructions, handling complex semantic and acoustic modifications without manual region specification.\n\n- 🔥 First unified continuous speech tokenizer for both understanding and generatio"},{"ref":"P24","kind":"page","title":"inclusionAI/Ming-UniVision repository metadata","date":"2026-06-11T03:17:59.449184+00:00","date_source":null,"source_url":"https://github.com/inclusionAI/Ming-UniVision","signal_url":null,"signal_json_url":null,"text":"# inclusionAI/Ming-UniVision\n\nDescription: Code release for Ming-UniVision: Joint Image Understanding and Geneation with a Continuous Unified Tokenizer\n\nLanguage: Python\n\nLicense: MIT\n\nStars: 143\n\nForks: 5\n\nOpen issues: 4\n\nCreated: 2025-09-30T11:36:01Z\n\nPushed: 2025-10-14T13:38:52Z\n\nDefault branch: main\n\nFork: no\n\nArchived: no\n\nREADME:\n# Ming-UniVision: Joint Image Understanding and Geneation with a Continuous Unified Tokenizer\n\n<p align=\"center\">\n<img src=\"./figures/ant-bailing.png\" width=\"100\"/>\n<p>\n\n<p align=\"center\">📄 <a href=\"https://arxiv.org/pdf/2510.06590\">Technical Report</a> | 📖<a href=\"https://inclusionai.github.io/blog/mingtok/\">Project Page</a> ｜🤗 <a href=\"https://huggingface.co/inclusionAI/Ming-UniVision-16B-A3B\">Hugging Face</a>｜ 🤖 <a href=\"https://www.modelscope.cn/models/inclusionAI/Ming-UniVision-16B-A3B\">ModelScope</a></p>\n\n## 🌍 Introduction\n\n🌐 Ming-UniVision is a groundbreaking multimodal large language model (MLLM) that unifies vision understanding, generation, and editing within a single autoregressive next-token prediction (NTP) framework, powered by MingTok — the first continuous, unified visual tokenizer. By eliminating discrete quantization and leveraging a shared continuous latent space, Ming-UniVision enables seamless, end-to-end multimodal reasoning across diverse tasks.\nTrained on high-fidelity continuous visual representations, Ming-UniVision supports multi-round, in-context vision-language interactions, such as iterative question answering, image generation, and semantic editing — all without needing to decode intermediate states into pixels. This enables efficient, coherent, and human-like multimodal dialogue with consistent feature dynamics throughout.\n\n- 🌐 **First NTP MLLM with Continuous Unified Vision Representations**: [Ming-UniVision](https://huggingface.co/inclusionAI/Ming-UniVision-16B-A3B)\nunifies vision and language via next-token prediction using continuous visual tokens — no discrete quantization, full autoregressive generative paradigm, and support for both understanding and generation in a shared latent space.\n- 🖼️ **First Continuous Unified Visual Tokenizer:** [MingTok-Vision](https://huggingface.co/inclusi"},{"ref":"P25","kind":"page","title":"inclusionAI/dInfer repository metadata","date":"2026-06-11T03:17:59.352678+00:00","date_source":null,"source_url":"https://github.com/inclusionAI/dInfer","signal_url":null,"signal_json_url":null,"text":"# inclusionAI/dInfer\n\nDescription: dInfer: An Efficient Inference Framework for Diffusion Language Models\n\nLanguage: Python\n\nLicense: Apache-2.0\n\nStars: 470\n\nForks: 45\n\nOpen issues: 18\n\nCreated: 2025-09-29T08:07:23Z\n\nPushed: 2026-02-11T03:10:42Z\n\nDefault branch: master\n\nFork: no\n\nArchived: no\n\nREADME:\n<div align=\"center\">\n<img src=\"assets/logo.svg\" width=\"40%\" alt=\"dInfer\" />\n</div>\n\n<h4 align=\"center\">\n\n[![License: MIT](https://img.shields.io/badge/License-Apache%202.0-blue.svg)](./LICENSE)\n[![HuggingFace: Models](https://img.shields.io/badge/HuggingFace-Models-yellow)](https://huggingface.co/inclusionAI/LLaDA-MoE-7B-A1B-Instruct)\n[![Technical Report: Arxiv](https://img.shields.io/badge/Technical%20Report-Arxiv-red)](https://arxiv.org/abs/2510.08666)\n\n<!-- [![arXiv][arxiv-image]][arxiv-url] -->\n\n</h4>\n\n## Introduction\ndInfer is an efficient and extensible inference framework for dLLMs. As illustrated in the following architecture, it modularizes inference into four components:\n*model*, *diffusion iteration manager*, *decoder* and *KV-cache manager*. It provides well-designed APIs for\nflexible algorithms combinations in each component. It now supports batched inference for improved throughput.\n\n<p align=\"center\">\n<img src=\"assets/Framework2.png\" alt=\"dInfer v0.1 architecture\" width=\"600\">\n<br>\n<b>Figure</b>: Overall Architecture of dInfer\n</p>\n\ndInfer supports multiple dLLM variants, including LLaDA, LLaDA-MoE and LLaDA2.\n\n## News\n**\\[2025/12/21\\]** release v0.2. The major features of this release can be found [here](https://github.com/inclusionAI/dInfer/releases/tag/v0.2.0).\n\n**\\[2025/12/10\\]** Support and speed up the formal version of block diffusion LLMs (LLaDA2-mini and LLaDA2-flash). Support quant versions of LLaDA2-mini and LLaDA2-flash.\n\n**\\[2025/11/15\\]** Support the inference on block diffusion LLMs (LLaDA2-mini-preview and LLaDA2-flash-preview).\n\n**\\[2025/10/10\\]** Release the first version of the dInfer framework.\n\n## Contents\n- [Supported Models](#supported-models)\n- [Quick Start](#quick-start)\n- [Benchmark Results](#benchmark-results)\n\n## Supported Models\n\ndInfer supports multiple diffusion language model variants with different architectures and s"},{"ref":"P26","kind":"page","title":"inclusionAI/Ming-VideoMAR repository metadata","date":"2026-06-11T03:17:59.100026+00:00","date_source":null,"source_url":"https://github.com/inclusionAI/Ming-VideoMAR","signal_url":null,"signal_json_url":null,"text":"# inclusionAI/Ming-VideoMAR\n\nLanguage: Python\n\nStars: 9\n\nForks: 1\n\nOpen issues: 0\n\nCreated: 2025-10-10T02:20:50Z\n\nPushed: 2025-10-22T08:43:25Z\n\nDefault branch: main\n\nFork: no\n\nArchived: no\n\nREADME:\n# Ming-VideoMAR: Autoregressive Video Generation with Continuous Tokens\n\n<p align=\"center\">\n<img src=\"./figures/ant-bailing.png\" width=\"100\"/>\n<p>\n\n<p align=\"center\">🤗 <a href=\"https://huggingface.co/inclusionAI/Ming-VideoMAR\">Hugging Face </a>｜📄 <a href=\"https://www.arxiv.org/abs/2506.14168\">Paper (NeurIPS 2025) </a> </p>\n\n## 🌍 Introduction\n\n- 🌐 **The First NTP MLLM with Continuous Unified Vision Representations:**\nMing-VideoMAR is a concise and efficient decoder-only autoregressive image-to-video model with continuous tokens, composing temporal frame-by-frame and spatial masked generation. Ming-VideoMAR identifies temporal causality and spatial bi-directionality as the first principle of video AR models, and proposes the next-frame diffusion loss for the integration of mask and video generation. \n- 🖼️ **First Zero-shot Resolution Scaling for Video Generation:** \nMing-VideoMAR replicates the unique capacity of sequence extrapolation from language models to video generation. It supports generating videos of flexible spatial and temporal resolutions that is far beyond the training resolution. This is achieved by solving the training-inference gap and adopting the 3D rotary embeddings. \n- ⚡ **Extreme Hihg Training Efficiency:**\nMing-VideoMAR proposes the temporal short-to-long curriculum learning and spatial progressive resolution training. It surpasses the previous state-of-the-art (Cosmos I2V) while requiring significantly fewer parameters (9.3%), training data (0.5%), and GPU resources (0.2%), both quantatively and qualitatively.\n- ⚡ **Extreme Hihg Inference Efficiency:**\nMing-VideoMAR inherently bears high efficiency due to simultaneous temporal-wise KV cache and spatial-wise parallel generation, significantly surpassing the NTP counterpart.\n- 🔗 **Accumulation Error Solution:**\nMing-VideoMAR employs the progressive temperature strategy at inference time to mitigate the accumulation error.\n\n<p align=\"center\">\n<img src=\"./figures/videomar-overall.png\" width=\"800"},{"ref":"P27","kind":"page","title":"inclusionAI/SWE-CARE repository metadata","date":"2026-06-11T03:17:59.082976+00:00","date_source":null,"source_url":"https://github.com/inclusionAI/SWE-CARE","signal_url":null,"signal_json_url":null,"text":"# inclusionAI/SWE-CARE\n\nLanguage: Python\n\nLicense: Apache-2.0\n\nStars: 15\n\nForks: 13\n\nOpen issues: 0\n\nCreated: 2025-10-10T07:36:39Z\n\nPushed: 2026-04-21T07:29:00Z\n\nDefault branch: main\n\nFork: no\n\nArchived: no\n\nREADME:\n# SWE-CARE: A Comprehensiveness-aware Benchmark for Evaluation of Code Review\n\n<p align=\"center\">\n<a href=\"https://arxiv.org/pdf/2509.14856\">\n<img src=\"https://img.shields.io/badge/Tech Report-arXiv-red\"></a>\n<a href=\"https://huggingface.co/datasets/inclusionAI/SWE-CARE\">\n<img src=\"https://img.shields.io/badge/Dataset-HuggingFace-orange\"></a>\n<a href=\"https://github.com/inclusionAI/SWE-CARE/blob/main/LICENSE\">\n<img src=\"https://img.shields.io/badge/License-Apache-blue\"></a>\n</p>\n\nA comprehensiveness-aware benchmark for repository-level CR evaluation.\n\n## 📝 Overview\n\nCode review (CR) refers to the process of having other developers on the team check the code written by a particular developer. It aims to improve the code quality and find code defects and plays an important role in software quality maintenance. Some research had proposed some CR benchmarks and automatic CR approaches. However, existing CR benchmarks and approaches, lack of comprehensiveness, which is not close to the real scenario. The rapid growth of Large Language Model (LLM) capabilities has made comprehensive CR a possibility. To evaluate the LLMs' performance in comprehensive CR, we construct a comprehensiveness-aware CR dataset in Python, namely SWE-CARE. The dataset is categorized into nine types and each instance's information covers the full process of code review. In addition, the repository-level feature is also included in each instance. Based on the dataset, we design a framework to evaluate LLM’s performance on CR.\n\n## 🛠️ Set Up\n\nFollow these steps to set up the project locally.\n\n1. **Clone the repository:**\n\n```bash\ngit clone https://github.com/your-username/SWE-CARE.git\ncd SWE-CARE\n```\n\n2. **Install dependencies:**\nThis project uses `uv` for package management. Make sure you have Python 3.10 or higher.\n\n```bash\npip install uv\nuv sync\n```\n\nAlternatively, you can use `pip`:\n\n```bash\npip install -e .\n```\n\n3. **Set up pre-commit hooks (for development):**\nThis project uses"},{"ref":"P28","kind":"page","title":"inclusionAI/linghe repository metadata","date":"2026-06-11T03:17:58.911312+00:00","date_source":null,"source_url":"https://github.com/inclusionAI/linghe","signal_url":null,"signal_json_url":null,"text":"# inclusionAI/linghe\n\nDescription: A high-performance kernel library for LLM training\n\nLanguage: Python\n\nLicense: MIT\n\nStars: 80\n\nForks: 10\n\nOpen issues: 1\n\nCreated: 2025-10-14T09:50:16Z\n\nPushed: 2026-04-28T09:54:12Z\n\nDefault branch: main\n\nFork: no\n\nArchived: no\n\nREADME:\n<h1 align=\"center\"> linghe </h1>\n\n<div style=\"text-align: center;\">\n<img src=\"assets/linghe.png\" alt=\"Logo\" width=\"200\">\n</div>\n\n<p align=\"center\">\nA library of high-performance kernels for LLM training.\n</p>\n\n## Roadmap ##\n---\n\n- Support more shapes and various GPU archs.\n- Release our fp8 training kernels beyond blockwise quantization.\n\n## *News or Update* 🔥\n---\n- [2025/07] We implement multiple kernels for FP8 training with `Megatron-LM` blockwise quantization. \n\n## Introduction\n---\nOur repo, linghe, is designed for LLM training, especially for MoE training with FP8 quantizaiton. It provides 3 main categories of kernels:\n\n- **Fused quantization kernels**: fuse quantization with previous layer, e.g., RMS norm and Silu.\n- **Memory-efficiency kernels**: fuse multiple IO-itensive operations, e.g., ROPE with qk-norm.\n- **Implementation-optimized kernels**: use efficient triton implementation, e.g., routing map padding instead of activation padding.\n\n## Benchmark\n---\nWe benchmark on H800 with batch size 8192, hidden size 2048, num experts 256, activation experts 8.\n\n| kernel | baseline(us) | linghe(us) | speedup |\n|--------|--------------|------------|---------|\n| RMSNorm+Quantization(forward) | 159.3 us | 72.4 us | 2.2 |\n| Split+qk-norm+rope+transpose(forward) | 472 us | 59.1 us | 7.99 |\n| Split+qk-norm+rope+transpose(backward) | 645 us | 107.5 us | 6.0 |\n| Fp32 router gemm(forward) | 242.3 us | 61.6 us | 3.931 |\n| Fp32 router gemm(backward) | 232.7 us | 78.1 us | 2.979 |\n| Permute with padded indices | 388 us | 229.4 us | 1.69 |\n| Unpermute with padding indices | 988.6 us | 806.9 us | 1.23 |\n| Batch Silu+quantization(forward) | 6241.7 us | 1181.7 us | 5.28 |\n| Batch Silu+quantization(backward) | 7147.7 us | 2317.9 us | 3.08 |\n| Silu+quantization(forward) | 144.9 us | 58.2 us | 2.48 |\n| Silu+quantization(backward) | 163.4 us | 74.2 us | 2.2 |\n| fused linear gate(forward) | 160.4 us | 46.9 us | 3."},{"ref":"E1","kind":"event","title":"inclusionAI/Ling-2.6-flash","date":"2026-04-28T03:27:56+00:00","date_source":"source","source_url":"https://huggingface.co/inclusionAI/Ling-2.6-flash","signal_url":"https://onlylabs.fyi/signals/5f271d7b-0454-436e-b047-20873d891685","signal_json_url":"https://onlylabs.fyi/signals/5f271d7b-0454-436e-b047-20873d891685/signal.json","text":"model_released · inclusionAI/Ling-2.6-flash · signal_desk=releases · occurred_at=2026-04-28T03:27:56+00:00 · url=https://huggingface.co/inclusionAI/Ling-2.6-flash · hf_downloads=9798 · hf_likes=493 · hf_params=107494409216 · pipeline=text-generation · license=mit"},{"ref":"E2","kind":"event","title":"inclusionAI/Ling-2.6-1T","date":"2026-04-29T03:19:36+00:00","date_source":"source","source_url":"https://huggingface.co/inclusionAI/Ling-2.6-1T","signal_url":"https://onlylabs.fyi/signals/d56c53e5-d719-424a-964b-24995a3277c7","signal_json_url":"https://onlylabs.fyi/signals/d56c53e5-d719-424a-964b-24995a3277c7/signal.json","text":"model_released · inclusionAI/Ling-2.6-1T · signal_desk=releases · occurred_at=2026-04-29T03:19:36+00:00 · url=https://huggingface.co/inclusionAI/Ling-2.6-1T · hf_downloads=1105 · hf_likes=469 · hf_params=1025657871744 · pipeline=text-generation · license=mit"},{"ref":"E3","kind":"event","title":"inclusionAI/LLaDA2.0-Uni","date":"2026-04-22T17:10:47+00:00","date_source":"source","source_url":"https://huggingface.co/inclusionAI/LLaDA2.0-Uni","signal_url":"https://onlylabs.fyi/signals/34bceb00-128e-426c-9c45-ec7f6b585baa","signal_json_url":"https://onlylabs.fyi/signals/34bceb00-128e-426c-9c45-ec7f6b585baa/signal.json","text":"model_released · inclusionAI/LLaDA2.0-Uni · signal_desk=releases · occurred_at=2026-04-22T17:10:47+00:00 · url=https://huggingface.co/inclusionAI/LLaDA2.0-Uni · hf_downloads=7218 · hf_likes=247 · hf_params=16322752256 · pipeline=any-to-any · license=apache-2.0 · raw={\"derived_reason\":\"first-party-finetune\"}"},{"ref":"E4","kind":"event","title":"inclusionAI/Ring-2.5-1T","date":"2026-02-10T12:45:40+00:00","date_source":"source","source_url":"https://huggingface.co/inclusionAI/Ring-2.5-1T","signal_url":"https://onlylabs.fyi/signals/d5e38bc2-bf43-4d87-a002-b704a7ae8e86","signal_json_url":"https://onlylabs.fyi/signals/d5e38bc2-bf43-4d87-a002-b704a7ae8e86/signal.json","text":"model_released · inclusionAI/Ring-2.5-1T · signal_desk=releases · occurred_at=2026-02-10T12:45:40+00:00 · url=https://huggingface.co/inclusionAI/Ring-2.5-1T · hf_downloads=33299 · hf_likes=245 · hf_params=1012474606720 · pipeline=text-generation · license=mit"},{"ref":"E5","kind":"event","title":"inclusionAI/TwinFlow-Z-Image-Turbo","date":"2025-12-23T08:42:21+00:00","date_source":"source","source_url":"https://huggingface.co/inclusionAI/TwinFlow-Z-Image-Turbo","signal_url":"https://onlylabs.fyi/signals/b0dbc6eb-a0a7-4b0c-bf7a-1443ef794625","signal_json_url":"https://onlylabs.fyi/signals/b0dbc6eb-a0a7-4b0c-bf7a-1443ef794625/signal.json","text":"model_released · inclusionAI/TwinFlow-Z-Image-Turbo · signal_desk=releases · occurred_at=2025-12-23T08:42:21+00:00 · url=https://huggingface.co/inclusionAI/TwinFlow-Z-Image-Turbo · hf_downloads=12 · hf_likes=212 · pipeline=text-to-image · license=apache-2.0 · raw={\"derived_reason\":\"first-party-finetune\"}"},{"ref":"E6","kind":"event","title":"inclusionAI/Ring-2.6-1T","date":"2026-05-14T08:10:02+00:00","date_source":"source","source_url":"https://huggingface.co/inclusionAI/Ring-2.6-1T","signal_url":"https://onlylabs.fyi/signals/dde86c8e-76c3-4d1e-83d6-49cf5c3f4ffb","signal_json_url":"https://onlylabs.fyi/signals/dde86c8e-76c3-4d1e-83d6-49cf5c3f4ffb/signal.json","text":"model_released · inclusionAI/Ring-2.6-1T · signal_desk=releases · occurred_at=2026-05-14T08:10:02+00:00 · url=https://huggingface.co/inclusionAI/Ring-2.6-1T · hf_downloads=4780 · hf_likes=99 · hf_params=1025657871744 · pipeline=text-generation · license=mit"},{"ref":"E7","kind":"event","title":"inclusionAI/LLaDA2.1-mini","date":"2026-02-09T07:27:07+00:00","date_source":"source","source_url":"https://huggingface.co/inclusionAI/LLaDA2.1-mini","signal_url":"https://onlylabs.fyi/signals/6c465dc0-9f4d-4682-906a-2fbdc717e82e","signal_json_url":"https://onlylabs.fyi/signals/6c465dc0-9f4d-4682-906a-2fbdc717e82e/signal.json","text":"model_released · inclusionAI/LLaDA2.1-mini · signal_desk=releases · occurred_at=2026-02-09T07:27:07+00:00 · url=https://huggingface.co/inclusionAI/LLaDA2.1-mini · hf_downloads=15190 · hf_likes=125 · hf_params=16255643392 · pipeline=text-generation · license=apache-2.0"},{"ref":"E8","kind":"event","title":"inclusionAI/Ling-2.5-1T","date":"2026-02-14T07:48:42+00:00","date_source":"source","source_url":"https://huggingface.co/inclusionAI/Ling-2.5-1T","signal_url":"https://onlylabs.fyi/signals/dfe64205-f04e-4f74-863b-a05df2cdd35a","signal_json_url":"https://onlylabs.fyi/signals/dfe64205-f04e-4f74-863b-a05df2cdd35a/signal.json","text":"model_released · inclusionAI/Ling-2.5-1T · signal_desk=releases · occurred_at=2026-02-14T07:48:42+00:00 · url=https://huggingface.co/inclusionAI/Ling-2.5-1T · hf_downloads=215 · hf_likes=115 · hf_params=1012231160320 · pipeline=text-generation · license=mit"},{"ref":"E9","kind":"event","title":"inclusionAI/LLaDA2.1-flash","date":"2026-02-09T10:47:26+00:00","date_source":"source","source_url":"https://huggingface.co/inclusionAI/LLaDA2.1-flash","signal_url":"https://onlylabs.fyi/signals/418b0cf0-a450-4f45-9912-fdf12c41065d","signal_json_url":"https://onlylabs.fyi/signals/418b0cf0-a450-4f45-9912-fdf12c41065d/signal.json","text":"model_released · inclusionAI/LLaDA2.1-flash · signal_desk=releases · occurred_at=2026-02-09T10:47:26+00:00 · url=https://huggingface.co/inclusionAI/LLaDA2.1-flash · hf_downloads=151361 · hf_likes=92 · hf_params=102889705216 · pipeline=text-generation · license=apache-2.0"},{"ref":"E10","kind":"event","title":"inclusionAI/ZwZ-8B","date":"2026-02-12T08:52:34+00:00","date_source":"source","source_url":"https://huggingface.co/inclusionAI/ZwZ-8B","signal_url":"https://onlylabs.fyi/signals/b251bbec-4fbe-420b-9108-81b86fd0c69e","signal_json_url":"https://onlylabs.fyi/signals/b251bbec-4fbe-420b-9108-81b86fd0c69e/signal.json","text":"model_released · inclusionAI/ZwZ-8B · signal_desk=releases · occurred_at=2026-02-12T08:52:34+00:00 · url=https://huggingface.co/inclusionAI/ZwZ-8B · hf_downloads=252 · hf_likes=45 · hf_params=8767123696 · pipeline=image-text-to-text · license=apache-2.0 · raw={\"derived_reason\":\"first-party-finetune\"}"},{"ref":"E11","kind":"event","title":"inclusionAI/humming v0.1.5","date":"2026-06-10T02:14:34+00:00","date_source":"source","source_url":"https://github.com/inclusionAI/humming/releases/tag/v0.1.5","signal_url":"https://onlylabs.fyi/signals/d04c3902-289e-4119-a160-8eacf1c58123","signal_json_url":"https://onlylabs.fyi/signals/d04c3902-289e-4119-a160-8eacf1c58123/signal.json","text":"release · inclusionAI/humming v0.1.5 · signal_desk=releases · occurred_at=2026-06-10T02:14:34+00:00 · url=https://github.com/inclusionAI/humming/releases/tag/v0.1.5 · raw={\"repo\":\"inclusionAI/humming\"}"},{"ref":"E12","kind":"event","title":"inclusionAI/humming v0.1.4","date":"2026-06-04T03:23:35+00:00","date_source":"source","source_url":"https://github.com/inclusionAI/humming/releases/tag/v0.1.4","signal_url":"https://onlylabs.fyi/signals/3c935802-6849-4ef0-9376-fce649f0f1f6","signal_json_url":"https://onlylabs.fyi/signals/3c935802-6849-4ef0-9376-fce649f0f1f6/signal.json","text":"release · inclusionAI/humming v0.1.4 · signal_desk=releases · occurred_at=2026-06-04T03:23:35+00:00 · url=https://github.com/inclusionAI/humming/releases/tag/v0.1.4 · raw={\"repo\":\"inclusionAI/humming\"}"},{"ref":"E13","kind":"event","title":"inclusionAI/humming v0.1.3","date":"2026-06-02T15:03:43+00:00","date_source":"source","source_url":"https://github.com/inclusionAI/humming/releases/tag/v0.1.3","signal_url":"https://onlylabs.fyi/signals/72ceb215-533e-4f8c-84cb-52fbbe0f553d","signal_json_url":"https://onlylabs.fyi/signals/72ceb215-533e-4f8c-84cb-52fbbe0f553d/signal.json","text":"release · inclusionAI/humming v0.1.3 · signal_desk=releases · occurred_at=2026-06-02T15:03:43+00:00 · url=https://github.com/inclusionAI/humming/releases/tag/v0.1.3 · raw={\"repo\":\"inclusionAI/humming\"}"},{"ref":"E14","kind":"event","title":"inclusionAI/UI-Venus-1.5-2B","date":"2026-02-09T07:46:22+00:00","date_source":"source","source_url":"https://huggingface.co/inclusionAI/UI-Venus-1.5-2B","signal_url":"https://onlylabs.fyi/signals/713ba9a0-db11-48ff-954d-68885a1fe1c3","signal_json_url":"https://onlylabs.fyi/signals/713ba9a0-db11-48ff-954d-68885a1fe1c3/signal.json","text":"model_released · inclusionAI/UI-Venus-1.5-2B · signal_desk=releases · occurred_at=2026-02-09T07:46:22+00:00 · url=https://huggingface.co/inclusionAI/UI-Venus-1.5-2B · hf_downloads=1769 · hf_likes=38 · hf_params=2438696960 · pipeline=image-text-to-text · license=apache-2.0"},{"ref":"E15","kind":"event","title":"inclusionAI/ARGenSeg-8B","date":"2026-05-15T02:33:52+00:00","date_source":"source","source_url":"https://huggingface.co/inclusionAI/ARGenSeg-8B","signal_url":"https://onlylabs.fyi/signals/107f41df-b146-4e20-b2f3-fe20b0beba2f","signal_json_url":"https://onlylabs.fyi/signals/107f41df-b146-4e20-b2f3-fe20b0beba2f/signal.json","text":"model_released · inclusionAI/ARGenSeg-8B · signal_desk=releases · occurred_at=2026-05-15T02:33:52+00:00 · url=https://huggingface.co/inclusionAI/ARGenSeg-8B · hf_downloads=69 · hf_likes=6 · hf_params=8220903299"},{"ref":"E16","kind":"event","title":"inclusionAI/AWorld v0.3.2","date":"2026-05-25T09:51:12+00:00","date_source":"source","source_url":"https://github.com/inclusionAI/AWorld/releases/tag/v0.3.2","signal_url":"https://onlylabs.fyi/signals/9c608253-650a-497f-ae85-729d41ddeb7a","signal_json_url":"https://onlylabs.fyi/signals/9c608253-650a-497f-ae85-729d41ddeb7a/signal.json","text":"release · inclusionAI/AWorld v0.3.2 · signal_desk=releases · occurred_at=2026-05-25T09:51:12+00:00 · url=https://github.com/inclusionAI/AWorld/releases/tag/v0.3.2 · raw={\"repo\":\"inclusionAI/AWorld\"}"},{"ref":"E17","kind":"event","title":"inclusionAI/humming v0.1.2","date":"2026-05-23T16:17:38+00:00","date_source":"source","source_url":"https://github.com/inclusionAI/humming/releases/tag/v0.1.2","signal_url":"https://onlylabs.fyi/signals/3a060243-96a7-4e12-b8ce-31c0e56ed989","signal_json_url":"https://onlylabs.fyi/signals/3a060243-96a7-4e12-b8ce-31c0e56ed989/signal.json","text":"release · inclusionAI/humming v0.1.2 · signal_desk=releases · occurred_at=2026-05-23T16:17:38+00:00 · url=https://github.com/inclusionAI/humming/releases/tag/v0.1.2 · raw={\"repo\":\"inclusionAI/humming\"}"},{"ref":"E18","kind":"event","title":"inclusionAI/humming v0.1.1","date":"2026-05-22T08:31:40+00:00","date_source":"source","source_url":"https://github.com/inclusionAI/humming/releases/tag/v0.1.1","signal_url":"https://onlylabs.fyi/signals/d949fd54-db49-49dd-86cd-3343fda99893","signal_json_url":"https://onlylabs.fyi/signals/d949fd54-db49-49dd-86cd-3343fda99893/signal.json","text":"release · inclusionAI/humming v0.1.1 · signal_desk=releases · occurred_at=2026-05-22T08:31:40+00:00 · url=https://github.com/inclusionAI/humming/releases/tag/v0.1.1 · raw={\"repo\":\"inclusionAI/humming\"}"},{"ref":"E19","kind":"event","title":"inclusionAI/AEnvironment v0.1.7","date":"2026-05-21T15:20:21+00:00","date_source":"source","source_url":"https://github.com/inclusionAI/AEnvironment/releases/tag/v0.1.7","signal_url":"https://onlylabs.fyi/signals/eff512d0-2157-4aef-ac6b-6fa4c907f755","signal_json_url":"https://onlylabs.fyi/signals/eff512d0-2157-4aef-ac6b-6fa4c907f755/signal.json","text":"release · inclusionAI/AEnvironment v0.1.7 · signal_desk=releases · occurred_at=2026-05-21T15:20:21+00:00 · url=https://github.com/inclusionAI/AEnvironment/releases/tag/v0.1.7 · raw={\"repo\":\"inclusionAI/AEnvironment\"}"},{"ref":"E20","kind":"event","title":"inclusionAI/ZwZ-4B","date":"2026-02-12T08:53:01+00:00","date_source":"source","source_url":"https://huggingface.co/inclusionAI/ZwZ-4B","signal_url":"https://onlylabs.fyi/signals/2e917e1d-eafb-46dd-a9b0-c15bdd566ce2","signal_json_url":"https://onlylabs.fyi/signals/2e917e1d-eafb-46dd-a9b0-c15bdd566ce2/signal.json","text":"model_released · inclusionAI/ZwZ-4B · signal_desk=releases · occurred_at=2026-02-12T08:53:01+00:00 · url=https://huggingface.co/inclusionAI/ZwZ-4B · hf_downloads=358 · hf_likes=33 · hf_params=4826771968 · pipeline=image-text-to-text · license=apache-2.0 · raw={\"derived_reason\":\"first-party-finetune\"}"},{"ref":"E21","kind":"event","title":"inclusionAI/DR-Venus-4B-RL","date":"2026-04-21T08:05:38+00:00","date_source":"source","source_url":"https://huggingface.co/inclusionAI/DR-Venus-4B-RL","signal_url":"https://onlylabs.fyi/signals/ff186850-cf17-4547-8a97-b1b67f8a35cc","signal_json_url":"https://onlylabs.fyi/signals/ff186850-cf17-4547-8a97-b1b67f8a35cc/signal.json","text":"model_released · inclusionAI/DR-Venus-4B-RL · signal_desk=releases · occurred_at=2026-04-21T08:05:38+00:00 · url=https://huggingface.co/inclusionAI/DR-Venus-4B-RL · hf_downloads=200 · hf_likes=13 · hf_params=4411424256"},{"ref":"E22","kind":"event","title":"inclusionAI/LLaDA2.0-Uni-FP8","date":"2026-05-06T08:03:43+00:00","date_source":"source","source_url":"https://huggingface.co/inclusionAI/LLaDA2.0-Uni-FP8","signal_url":"https://onlylabs.fyi/signals/d6c4a960-010a-4ca7-8e50-d0daac87bfaf","signal_json_url":"https://onlylabs.fyi/signals/d6c4a960-010a-4ca7-8e50-d0daac87bfaf/signal.json","text":"model_released · inclusionAI/LLaDA2.0-Uni-FP8 · signal_desk=releases · occurred_at=2026-05-06T08:03:43+00:00 · url=https://huggingface.co/inclusionAI/LLaDA2.0-Uni-FP8 · hf_downloads=6033 · hf_likes=5 · hf_params=16323686144 · pipeline=any-to-any · license=apache-2.0 · raw={\"derived_reason\":\"first-party-finetune\"}"},{"ref":"E23","kind":"event","title":"inclusionAI/humming v0.1.0","date":"2026-05-15T06:32:56+00:00","date_source":"source","source_url":"https://github.com/inclusionAI/humming/releases/tag/v0.1.0","signal_url":"https://onlylabs.fyi/signals/b7c9e272-2b24-4c10-91b0-c75a87278b5f","signal_json_url":"https://onlylabs.fyi/signals/b7c9e272-2b24-4c10-91b0-c75a87278b5f/signal.json","text":"release · inclusionAI/humming v0.1.0 · signal_desk=releases · occurred_at=2026-05-15T06:32:56+00:00 · url=https://github.com/inclusionAI/humming/releases/tag/v0.1.0 · raw={\"repo\":\"inclusionAI/humming\"}"},{"ref":"E24","kind":"event","title":"inclusionAI/ARGenSeg","date":"2026-05-14T15:36:42+00:00","date_source":"source","source_url":"https://github.com/inclusionAI/ARGenSeg","signal_url":"https://onlylabs.fyi/signals/275d30a7-26c0-4331-8e3b-fcc34b6af9ef","signal_json_url":"https://onlylabs.fyi/signals/275d30a7-26c0-4331-8e3b-fcc34b6af9ef/signal.json","text":"repo_new · inclusionAI/ARGenSeg · signal_desk=repos · occurred_at=2026-05-14T15:36:42+00:00 · url=https://github.com/inclusionAI/ARGenSeg · stars=8 · raw={\"repo\":\"inclusionAI/ARGenSeg\",\"description\":\"[NeurIPS 2025] ARGenSeg: Image Segmentation with Autoregressive Image Generation Model\",\"language\":\"Python\"}"},{"ref":"E25","kind":"event","title":"inclusionAI/UI-Venus-1.5-8B","date":"2026-02-09T07:50:09+00:00","date_source":"source","source_url":"https://huggingface.co/inclusionAI/UI-Venus-1.5-8B","signal_url":"https://onlylabs.fyi/signals/2e645e4a-7a46-4c1f-9eb4-9fe263538434","signal_json_url":"https://onlylabs.fyi/signals/2e645e4a-7a46-4c1f-9eb4-9fe263538434/signal.json","text":"model_released · inclusionAI/UI-Venus-1.5-8B · signal_desk=releases · occurred_at=2026-02-09T07:50:09+00:00 · url=https://huggingface.co/inclusionAI/UI-Venus-1.5-8B · hf_downloads=4895 · hf_likes=28 · hf_params=8767123696 · pipeline=image-text-to-text · license=apache-2.0"},{"ref":"E26","kind":"event","title":"inclusionAI/UI-Venus-1.5-30B-A3B","date":"2026-02-09T07:55:59+00:00","date_source":"source","source_url":"https://huggingface.co/inclusionAI/UI-Venus-1.5-30B-A3B","signal_url":"https://onlylabs.fyi/signals/a98c1ed4-7751-4380-bfe5-ec05f6442b14","signal_json_url":"https://onlylabs.fyi/signals/a98c1ed4-7751-4380-bfe5-ec05f6442b14/signal.json","text":"model_released · inclusionAI/UI-Venus-1.5-30B-A3B · signal_desk=releases · occurred_at=2026-02-09T07:55:59+00:00 · url=https://huggingface.co/inclusionAI/UI-Venus-1.5-30B-A3B · hf_downloads=1602 · hf_likes=28 · hf_params=31070754032 · pipeline=image-text-to-text · license=apache-2.0"},{"ref":"E27","kind":"event","title":"inclusionAI/DR-Venus-4B-SFT","date":"2026-04-21T07:29:33+00:00","date_source":"source","source_url":"https://huggingface.co/inclusionAI/DR-Venus-4B-SFT","signal_url":"https://onlylabs.fyi/signals/a3865d6a-62d0-4da5-bcd7-395f5b680bcf","signal_json_url":"https://onlylabs.fyi/signals/a3865d6a-62d0-4da5-bcd7-395f5b680bcf/signal.json","text":"model_released · inclusionAI/DR-Venus-4B-SFT · signal_desk=releases · occurred_at=2026-04-21T07:29:33+00:00 · url=https://huggingface.co/inclusionAI/DR-Venus-4B-SFT · hf_downloads=232 · hf_likes=7 · hf_params=4411424256"},{"ref":"E28","kind":"event","title":"inclusionAI/TC-AE","date":"2026-04-08T06:06:01+00:00","date_source":"source","source_url":"https://huggingface.co/inclusionAI/TC-AE","signal_url":"https://onlylabs.fyi/signals/d72f4b36-00f4-47a8-8baa-12c316363cd2","signal_json_url":"https://onlylabs.fyi/signals/d72f4b36-00f4-47a8-8baa-12c316363cd2/signal.json","text":"model_released · inclusionAI/TC-AE · signal_desk=releases · occurred_at=2026-04-08T06:06:01+00:00 · url=https://huggingface.co/inclusionAI/TC-AE · hf_likes=13 · pipeline=feature-extraction · license=mit"},{"ref":"E29","kind":"event","title":"inclusionAI/LLaDA2.0-Uni","date":"2026-04-17T08:48:24+00:00","date_source":"source","source_url":"https://github.com/inclusionAI/LLaDA2.0-Uni","signal_url":"https://onlylabs.fyi/signals/9a21494b-5ba2-424e-87c9-16abaf1356a3","signal_json_url":"https://onlylabs.fyi/signals/9a21494b-5ba2-424e-87c9-16abaf1356a3/signal.json","text":"repo_new · inclusionAI/LLaDA2.0-Uni · signal_desk=repos · occurred_at=2026-04-17T08:48:24+00:00 · url=https://github.com/inclusionAI/LLaDA2.0-Uni · stars=758 · raw={\"repo\":\"inclusionAI/LLaDA2.0-Uni\",\"description\":\"LLaDA2.0-Uni: Understanding and Generation the World. \",\"language\":\"Python\"}"},{"ref":"E30","kind":"event","title":"inclusionAI/asystem-awex v0.7.0","date":"2026-04-22T09:13:47+00:00","date_source":"source","source_url":"https://github.com/inclusionAI/asystem-awex/releases/tag/v0.7.0","signal_url":"https://onlylabs.fyi/signals/2b92f988-d412-4dd8-a48a-f2efbccc98fa","signal_json_url":"https://onlylabs.fyi/signals/2b92f988-d412-4dd8-a48a-f2efbccc98fa/signal.json","text":"release · inclusionAI/asystem-awex v0.7.0 · signal_desk=releases · occurred_at=2026-04-22T09:13:47+00:00 · url=https://github.com/inclusionAI/asystem-awex/releases/tag/v0.7.0 · raw={\"repo\":\"inclusionAI/asystem-awex\"}"},{"ref":"E31","kind":"event","title":"inclusionAI/DR-Venus","date":"2026-04-20T09:31:31+00:00","date_source":"source","source_url":"https://github.com/inclusionAI/DR-Venus","signal_url":"https://onlylabs.fyi/signals/1817732d-ce18-4e4f-a139-c3fb5e507a49","signal_json_url":"https://onlylabs.fyi/signals/1817732d-ce18-4e4f-a139-c3fb5e507a49/signal.json","text":"repo_new · inclusionAI/DR-Venus · signal_desk=repos · occurred_at=2026-04-20T09:31:31+00:00 · url=https://github.com/inclusionAI/DR-Venus · stars=83 · raw={\"repo\":\"inclusionAI/DR-Venus\",\"language\":\"Python\"}"},{"ref":"E32","kind":"event","title":"inclusionAI/asystem-awex v0.6.0","date":"2026-04-18T08:45:58+00:00","date_source":"source","source_url":"https://github.com/inclusionAI/asystem-awex/releases/tag/v0.6.0","signal_url":"https://onlylabs.fyi/signals/810407bc-fc08-4649-8e1d-f9de42e3fbb1","signal_json_url":"https://onlylabs.fyi/signals/810407bc-fc08-4649-8e1d-f9de42e3fbb1/signal.json","text":"release · inclusionAI/asystem-awex v0.6.0 · signal_desk=releases · occurred_at=2026-04-18T08:45:58+00:00 · url=https://github.com/inclusionAI/asystem-awex/releases/tag/v0.6.0 · raw={\"repo\":\"inclusionAI/asystem-awex\"}"},{"ref":"E33","kind":"event","title":"inclusionAI/asystem-awex v0.5.0","date":"2026-04-13T10:10:33+00:00","date_source":"source","source_url":"https://github.com/inclusionAI/asystem-awex/releases/tag/v0.5.0","signal_url":"https://onlylabs.fyi/signals/4d2ab62a-b381-45e2-9a17-8f3ec004c1dd","signal_json_url":"https://onlylabs.fyi/signals/4d2ab62a-b381-45e2-9a17-8f3ec004c1dd/signal.json","text":"release · inclusionAI/asystem-awex v0.5.0 · signal_desk=releases · occurred_at=2026-04-13T10:10:33+00:00 · url=https://github.com/inclusionAI/asystem-awex/releases/tag/v0.5.0 · raw={\"repo\":\"inclusionAI/asystem-awex\"}"},{"ref":"E34","kind":"event","title":"inclusionAI/asystem-awex v0.4.0","date":"2026-04-10T13:40:25+00:00","date_source":"source","source_url":"https://github.com/inclusionAI/asystem-awex/releases/tag/v0.4.0","signal_url":"https://onlylabs.fyi/signals/01c0721b-f125-4fec-8da6-cc51786688f7","signal_json_url":"https://onlylabs.fyi/signals/01c0721b-f125-4fec-8da6-cc51786688f7/signal.json","text":"release · inclusionAI/asystem-awex v0.4.0 · signal_desk=releases · occurred_at=2026-04-10T13:40:25+00:00 · url=https://github.com/inclusionAI/asystem-awex/releases/tag/v0.4.0 · raw={\"repo\":\"inclusionAI/asystem-awex\"}"},{"ref":"E35","kind":"event","title":"inclusionAI/TC-AE","date":"2026-04-07T09:26:43+00:00","date_source":"source","source_url":"https://github.com/inclusionAI/TC-AE","signal_url":"https://onlylabs.fyi/signals/de8b40c9-261c-48b6-a3c5-5012b2eb971b","signal_json_url":"https://onlylabs.fyi/signals/de8b40c9-261c-48b6-a3c5-5012b2eb971b/signal.json","text":"repo_new · inclusionAI/TC-AE · signal_desk=repos · occurred_at=2026-04-07T09:26:43+00:00 · url=https://github.com/inclusionAI/TC-AE · stars=25 · raw={\"repo\":\"inclusionAI/TC-AE\",\"description\":\"Official repo for \\\"TC-AE: Unlocking Token Capacity for Deep Compression Autoencoders\\\"\",\"language\":\"Python\"}"},{"ref":"E36","kind":"event","title":"inclusionAI/cuLA","date":"2026-04-02T07:05:32+00:00","date_source":"source","source_url":"https://github.com/inclusionAI/cuLA","signal_url":"https://onlylabs.fyi/signals/f4b1629a-a734-48e8-be6b-7f9e4b5525b6","signal_json_url":"https://onlylabs.fyi/signals/f4b1629a-a734-48e8-be6b-7f9e4b5525b6/signal.json","text":"repo_new · inclusionAI/cuLA · signal_desk=repos · occurred_at=2026-04-02T07:05:32+00:00 · url=https://github.com/inclusionAI/cuLA · stars=519 · raw={\"repo\":\"inclusionAI/cuLA\",\"description\":\"CUDA kernels for linear attention variants, written in CuTe DSL and CUTLASS C++.\",\"language\":\"Python\"}"},{"ref":"E37","kind":"event","title":"inclusionAI/Ming-omni-tts-tokenizer-12Hz","date":"2026-02-11T03:12:52+00:00","date_source":"source","source_url":"https://huggingface.co/inclusionAI/Ming-omni-tts-tokenizer-12Hz","signal_url":"https://onlylabs.fyi/signals/527f5ff3-cfcc-4384-ae72-4ceb68de1f49","signal_json_url":"https://onlylabs.fyi/signals/527f5ff3-cfcc-4384-ae72-4ceb68de1f49/signal.json","text":"model_released · inclusionAI/Ming-omni-tts-tokenizer-12Hz · signal_desk=releases · occurred_at=2026-02-11T03:12:52+00:00 · url=https://huggingface.co/inclusionAI/Ming-omni-tts-tokenizer-12Hz · hf_downloads=67 · hf_likes=12 · hf_params=780389010 · pipeline=audio-to-audio · license=apache-2.0"},{"ref":"E38","kind":"event","title":"inclusionAI/ZwZ-7B","date":"2026-02-12T08:52:48+00:00","date_source":"source","source_url":"https://huggingface.co/inclusionAI/ZwZ-7B","signal_url":"https://onlylabs.fyi/signals/86a0530b-7dcd-4789-9c96-67942f0ac9e2","signal_json_url":"https://onlylabs.fyi/signals/86a0530b-7dcd-4789-9c96-67942f0ac9e2/signal.json","text":"model_released · inclusionAI/ZwZ-7B · signal_desk=releases · occurred_at=2026-02-12T08:52:48+00:00 · url=https://huggingface.co/inclusionAI/ZwZ-7B · hf_downloads=26 · hf_likes=12 · hf_params=8292166656 · pipeline=image-text-to-text · license=apache-2.0 · raw={\"derived_reason\":\"first-party-finetune\"}"},{"ref":"E39","kind":"event","title":"inclusionAI/LLaDA2.0-mini-CAP","date":"2025-12-10T13:21:10+00:00","date_source":"source","source_url":"https://huggingface.co/inclusionAI/LLaDA2.0-mini-CAP","signal_url":"https://onlylabs.fyi/signals/bbe145c2-15d7-48ad-95c6-6dff18b5b294","signal_json_url":"https://onlylabs.fyi/signals/bbe145c2-15d7-48ad-95c6-6dff18b5b294/signal.json","text":"model_released · inclusionAI/LLaDA2.0-mini-CAP · signal_desk=releases · occurred_at=2025-12-10T13:21:10+00:00 · url=https://huggingface.co/inclusionAI/LLaDA2.0-mini-CAP · hf_downloads=155 · hf_likes=10 · hf_params=16255643392 · pipeline=text-generation · license=apache-2.0"},{"ref":"E40","kind":"event","title":"inclusionAI/LLaDA2.0-flash-CAP","date":"2025-12-10T12:20:56+00:00","date_source":"source","source_url":"https://huggingface.co/inclusionAI/LLaDA2.0-flash-CAP","signal_url":"https://onlylabs.fyi/signals/06c73646-e48b-4526-93ed-21e061a7dc76","signal_json_url":"https://onlylabs.fyi/signals/06c73646-e48b-4526-93ed-21e061a7dc76/signal.json","text":"model_released · inclusionAI/LLaDA2.0-flash-CAP · signal_desk=releases · occurred_at=2025-12-10T12:20:56+00:00 · url=https://huggingface.co/inclusionAI/LLaDA2.0-flash-CAP · hf_downloads=33 · hf_likes=10 · hf_params=102889705216 · pipeline=text-generation · license=apache-2.0"},{"ref":"E41","kind":"event","title":"inclusionAI/cuLA v0.1.0","date":"2026-04-02T10:43:29+00:00","date_source":"source","source_url":"https://github.com/inclusionAI/cuLA/releases/tag/v0.1.0","signal_url":"https://onlylabs.fyi/signals/60173cce-ee3a-4739-bd5c-5e00ec2586d0","signal_json_url":"https://onlylabs.fyi/signals/60173cce-ee3a-4739-bd5c-5e00ec2586d0/signal.json","text":"release · inclusionAI/cuLA v0.1.0 · signal_desk=releases · occurred_at=2026-04-02T10:43:29+00:00 · url=https://github.com/inclusionAI/cuLA/releases/tag/v0.1.0 · raw={\"repo\":\"inclusionAI/cuLA\"}"},{"ref":"E42","kind":"event","title":"Taking the Pulse of Agentic AI from the Developer Community at the End of Q1 2026","date":"2026-04-01T00:00:00+00:00","date_source":"rss.item_date","source_url":"https://www.inclusion-ai.org/blog/agentic-landscape-in-2026-Q1","signal_url":"https://onlylabs.fyi/signals/2cbb5136-91f1-45ad-ab49-648a7e5cdf57","signal_json_url":"https://onlylabs.fyi/signals/2cbb5136-91f1-45ad-ab49-648a7e5cdf57/signal.json","text":"post_published · Taking the Pulse of Agentic AI from the Developer Community at the End of Q1 2026 · signal_desk=talking · occurred_at=2026-04-01T00:00:00+00:00 · url=https://www.inclusion-ai.org/blog/agentic-landscape-in-2026-Q1 · raw={\"excerpt\":\"Today, I want to share some observations on the Agentic AI ecosystem from the vantage point of 2026's first quarter—technical trends read from popular projects, portraits of AI developers, and the subtle relationship between developers and AI tools. This is not meant to be comprehensive; we welcome the community to share more observations and reflections.\"}"},{"ref":"E43","kind":"event","title":"inclusionAI/ZwZ-2B","date":"2026-03-04T08:25:43+00:00","date_source":"source","source_url":"https://huggingface.co/inclusionAI/ZwZ-2B","signal_url":"https://onlylabs.fyi/signals/14f5aea9-1306-4ef0-9771-2db1dc46ef18","signal_json_url":"https://onlylabs.fyi/signals/14f5aea9-1306-4ef0-9771-2db1dc46ef18/signal.json","text":"model_released · inclusionAI/ZwZ-2B · signal_desk=releases · occurred_at=2026-03-04T08:25:43+00:00 · url=https://huggingface.co/inclusionAI/ZwZ-2B · hf_downloads=920 · hf_likes=6 · hf_params=2438696960 · pipeline=image-text-to-text · license=apache-2.0 · raw={\"derived_reason\":\"first-party-finetune\"}"},{"ref":"E44","kind":"event","title":"inclusionAI/AWorld","date":"2025-03-14T08:30:52+00:00","date_source":"source","source_url":"https://github.com/inclusionAI/AWorld","signal_url":"https://onlylabs.fyi/signals/a81bfcf6-c148-4a41-ba3f-96a4acfaa7c8","signal_json_url":"https://onlylabs.fyi/signals/a81bfcf6-c148-4a41-ba3f-96a4acfaa7c8/signal.json","text":"repo_new · inclusionAI/AWorld · signal_desk=repos · occurred_at=2025-03-14T08:30:52+00:00 · url=https://github.com/inclusionAI/AWorld · stars=1202 · raw={\"repo\":\"inclusionAI/AWorld\",\"description\":\"Search, understand, reproduce, and improve an idea with ease\",\"language\":\"Python\"}"},{"ref":"E45","kind":"event","title":"inclusionAI/AReaL-SEA-235B-A22B","date":"2026-03-01T09:56:04+00:00","date_source":"source","source_url":"https://huggingface.co/inclusionAI/AReaL-SEA-235B-A22B","signal_url":"https://onlylabs.fyi/signals/6d1d1b59-36e4-47c9-a8f7-4d3ea23813cf","signal_json_url":"https://onlylabs.fyi/signals/6d1d1b59-36e4-47c9-a8f7-4d3ea23813cf/signal.json","text":"model_released · inclusionAI/AReaL-SEA-235B-A22B · signal_desk=releases · occurred_at=2026-03-01T09:56:04+00:00 · url=https://huggingface.co/inclusionAI/AReaL-SEA-235B-A22B · hf_downloads=25 · hf_likes=6 · hf_params=235093634560 · pipeline=text-generation · license=apache-2.0 · raw={\"derived_reason\":\"first-party-finetune\"}"},{"ref":"E46","kind":"event","title":"inclusionAI/UI-Venus","date":"2025-08-15T11:57:47+00:00","date_source":"source","source_url":"https://github.com/inclusionAI/UI-Venus","signal_url":"https://onlylabs.fyi/signals/15f91882-04b1-47b3-b041-f591d9f88dd6","signal_json_url":"https://onlylabs.fyi/signals/15f91882-04b1-47b3-b041-f591d9f88dd6/signal.json","text":"repo_new · inclusionAI/UI-Venus · signal_desk=repos · occurred_at=2025-08-15T11:57:47+00:00 · url=https://github.com/inclusionAI/UI-Venus · stars=1010 · raw={\"repo\":\"inclusionAI/UI-Venus\",\"description\":\"UI-Venus is a native UI agent designed to perform precise GUI element grounding and effective navigation using only screenshots as input.\",\"language\":\"Python\"}"},{"ref":"E47","kind":"event","title":"inclusionAI/asystem-awex v0.3.0","date":"2026-03-23T06:55:41+00:00","date_source":"source","source_url":"https://github.com/inclusionAI/asystem-awex/releases/tag/v0.3.0","signal_url":"https://onlylabs.fyi/signals/380416fa-511b-4952-895a-e26fd6b7e47b","signal_json_url":"https://onlylabs.fyi/signals/380416fa-511b-4952-895a-e26fd6b7e47b/signal.json","text":"release · inclusionAI/asystem-awex v0.3.0 · signal_desk=releases · occurred_at=2026-03-23T06:55:41+00:00 · url=https://github.com/inclusionAI/asystem-awex/releases/tag/v0.3.0 · raw={\"repo\":\"inclusionAI/asystem-awex\"}"},{"ref":"E48","kind":"event","title":"inclusionAI/Ming","date":"2025-04-21T07:39:03+00:00","date_source":"source","source_url":"https://github.com/inclusionAI/Ming","signal_url":"https://onlylabs.fyi/signals/4e634387-b24f-4ca6-bc57-34cc16764285","signal_json_url":"https://onlylabs.fyi/signals/4e634387-b24f-4ca6-bc57-34cc16764285/signal.json","text":"repo_new · inclusionAI/Ming · signal_desk=repos · occurred_at=2025-04-21T07:39:03+00:00 · url=https://github.com/inclusionAI/Ming · stars=656 · raw={\"repo\":\"inclusionAI/Ming\",\"description\":\"Ming - facilitating advanced multimodal understanding and generation capabilities built upon the Ling LLM.\",\"language\":\"Jupyter Notebook\"}"},{"ref":"E49","kind":"event","title":"inclusionAI/gorilla","date":"2025-06-09T07:23:19+00:00","date_source":"source","source_url":"https://github.com/inclusionAI/gorilla","signal_url":"https://onlylabs.fyi/signals/186776da-2254-4ba6-8534-e95164022130","signal_json_url":"https://onlylabs.fyi/signals/186776da-2254-4ba6-8534-e95164022130/signal.json","text":"repo_forked · inclusionAI/gorilla · signal_desk=forks · occurred_at=2025-06-09T07:23:19+00:00 · url=https://github.com/inclusionAI/gorilla · stars=2 · raw={\"repo\":\"inclusionAI/gorilla\",\"parent\":\"ShishirPatil/gorilla\"}"},{"ref":"E50","kind":"event","title":"Ming-Omni-TTS: Simple and Efficient Unified Generation of Speech, Music, and Sound with Precise Control","date":"2026-03-04T00:00:00+00:00","date_source":"rss.item_date","source_url":"https://www.inclusion-ai.org/blog/ming-omni-tts","signal_url":"https://onlylabs.fyi/signals/5e1cfc8a-4c7d-492f-a99f-dd64392daa52","signal_json_url":"https://onlylabs.fyi/signals/5e1cfc8a-4c7d-492f-a99f-dd64392daa52/signal.json","text":"post_published · Ming-Omni-TTS: Simple and Efficient Unified Generation of Speech, Music, and Sound with Precise Control · signal_desk=talking · occurred_at=2026-03-04T00:00:00+00:00 · url=https://www.inclusion-ai.org/blog/ming-omni-tts · raw={\"excerpt\":\"GITHUB 🤗 Hugging Face｜ 🤖 ModelScope\"}"},{"ref":"E51","kind":"event","title":"The Community Stories of vLLM and SGLang","date":"2025-12-17T00:00:00+00:00","date_source":"rss.item_date","source_url":"https://www.inclusion-ai.org/blog/llm-landscape-vllm-sgl","signal_url":"https://onlylabs.fyi/signals/957b8d9f-1548-4ee4-a149-d34b39ed4dea","signal_json_url":"https://onlylabs.fyi/signals/957b8d9f-1548-4ee4-a149-d34b39ed4dea/signal.json","text":"post_published · The Community Stories of vLLM and SGLang · signal_desk=talking · occurred_at=2025-12-17T00:00:00+00:00 · url=https://www.inclusion-ai.org/blog/llm-landscape-vllm-sgl · raw={\"excerpt\":\"Originally published on Medium by Ant Open Source.\"}"},{"ref":"E52","kind":"event","title":"Ming-flash-omni-Preview: A Sparse, Unified Architecture for Multimodal Perception and Generation","date":"2025-10-28T00:00:00+00:00","date_source":"rss.item_date","source_url":"https://www.inclusion-ai.org/blog/ming-flash-omni-preview","signal_url":"https://onlylabs.fyi/signals/12b75563-2758-4e3e-9e19-672d996b030c","signal_json_url":"https://onlylabs.fyi/signals/12b75563-2758-4e3e-9e19-672d996b030c/signal.json","text":"post_published · Ming-flash-omni-Preview: A Sparse, Unified Architecture for Multimodal Perception and Generation · signal_desk=talking · occurred_at=2025-10-28T00:00:00+00:00 · url=https://www.inclusion-ai.org/blog/ming-flash-omni-preview · raw={\"excerpt\":\"GITHUB ARXIV 🤗 Hugging Face｜ 🤖 ModelScope\"}"},{"ref":"E53","kind":"event","title":"Open Source LLM Development Landscape 2.0: 2025 Revisited","date":"2025-10-11T00:00:00+00:00","date_source":"rss.item_date","source_url":"https://www.inclusion-ai.org/blog/llm-landscape-2-0","signal_url":"https://onlylabs.fyi/signals/1fb31514-221f-4f68-bbdf-52eae4fa2366","signal_json_url":"https://onlylabs.fyi/signals/1fb31514-221f-4f68-bbdf-52eae4fa2366/signal.json","text":"post_published · Open Source LLM Development Landscape 2.0: 2025 Revisited · signal_desk=talking · occurred_at=2025-10-11T00:00:00+00:00 · url=https://www.inclusion-ai.org/blog/llm-landscape-2-0 · raw={\"excerpt\":\"Originally published on Medium by Ant Open Source.\"}"},{"ref":"E54","kind":"event","title":"Ming-UniAudio: Speech LLM for Joint Understanding, Generation and Editing with Unified Representation","date":"2025-10-01T00:00:00+00:00","date_source":"rss.item_date","source_url":"https://www.inclusion-ai.org/blog/ming-uniaudio","signal_url":"https://onlylabs.fyi/signals/6443bea9-11eb-40ed-9c4f-9648b55c8137","signal_json_url":"https://onlylabs.fyi/signals/6443bea9-11eb-40ed-9c4f-9648b55c8137/signal.json","text":"post_published · Ming-UniAudio: Speech LLM for Joint Understanding, Generation and Editing with Unified Representation · signal_desk=talking · occurred_at=2025-10-01T00:00:00+00:00 · url=https://www.inclusion-ai.org/blog/ming-uniaudio · raw={\"excerpt\":\"GITHUB 🤗 Hugging Face｜ 🤖 ModelScope\"}"},{"ref":"E55","kind":"event","title":"Ming-UniVision: Joint Image Understanding and Generation via a Unified Continuous Tokenizer","date":"2025-10-01T00:00:00+00:00","date_source":"rss.item_date","source_url":"https://www.inclusion-ai.org/blog/mingtok","signal_url":"https://onlylabs.fyi/signals/1863c4ea-a67d-4d7a-8a70-898ca2e4f924","signal_json_url":"https://onlylabs.fyi/signals/1863c4ea-a67d-4d7a-8a70-898ca2e4f924/signal.json","text":"post_published · Ming-UniVision: Joint Image Understanding and Generation via a Unified Continuous Tokenizer · signal_desk=talking · occurred_at=2025-10-01T00:00:00+00:00 · url=https://www.inclusion-ai.org/blog/mingtok · raw={\"excerpt\":\"GITHUB 🤗 Hugging Face｜ 🤖 ModelScope\"}"},{"ref":"E56","kind":"event","title":"Segmentation-as-Editing for Unified Multimodal AI","date":"2025-09-13T00:00:00+00:00","date_source":"rss.item_date","source_url":"https://www.inclusion-ai.org/blog/ming-lite-omni-1_5-seg","signal_url":"https://onlylabs.fyi/signals/b21843d1-15a8-4df9-9d04-cba656dfdfcf","signal_json_url":"https://onlylabs.fyi/signals/b21843d1-15a8-4df9-9d04-cba656dfdfcf/signal.json","text":"post_published · Segmentation-as-Editing for Unified Multimodal AI · signal_desk=talking · occurred_at=2025-09-13T00:00:00+00:00 · url=https://www.inclusion-ai.org/blog/ming-lite-omni-1_5-seg · raw={\"excerpt\":\"GITHUB 🤗 Hugging Face｜ 🤖 ModelScope\"}"},{"ref":"E57","kind":"event","title":"Introducing Ring-lite-2507","date":"2025-08-05T00:00:00+00:00","date_source":"rss.item_date","source_url":"https://www.inclusion-ai.org/blog/ring-lite-2507","signal_url":"https://onlylabs.fyi/signals/bea46404-1680-469a-956a-02cc4cb245da","signal_json_url":"https://onlylabs.fyi/signals/bea46404-1680-469a-956a-02cc4cb245da/signal.json","text":"post_published · Introducing Ring-lite-2507 · signal_desk=talking · occurred_at=2025-08-05T00:00:00+00:00 · url=https://www.inclusion-ai.org/blog/ring-lite-2507 · raw={\"excerpt\":\"📖 Technical Report | 🤗 Hugging Face｜ 🤖 ModelScope\"}"},{"ref":"E58","kind":"event","title":"Introducing Ming-Lite-Omni V1.5","date":"2025-07-21T00:00:00+00:00","date_source":"rss.item_date","source_url":"https://www.inclusion-ai.org/blog/ming-lite-omni-1_5","signal_url":"https://onlylabs.fyi/signals/bd0e16ae-5964-4310-bd66-038b5ba0189f","signal_json_url":"https://onlylabs.fyi/signals/bd0e16ae-5964-4310-bd66-038b5ba0189f/signal.json","text":"post_published · Introducing Ming-Lite-Omni V1.5 · signal_desk=talking · occurred_at=2025-07-21T00:00:00+00:00 · url=https://www.inclusion-ai.org/blog/ming-lite-omni-1_5 · raw={\"excerpt\":\"GITHUB 🤗 Hugging Face｜ 🤖 ModelScope\"}"},{"ref":"E59","kind":"event","title":"M2-Reasoning: Empowering MLLMs with Unified General and Spatial Reasoning","date":"2025-07-11T00:00:00+00:00","date_source":"rss.item_date","source_url":"https://www.inclusion-ai.org/blog/m2-reasoning","signal_url":"https://onlylabs.fyi/signals/b038ba8b-9f91-4768-96c4-5bfa672a3553","signal_json_url":"https://onlylabs.fyi/signals/b038ba8b-9f91-4768-96c4-5bfa672a3553/signal.json","text":"post_published · M2-Reasoning: Empowering MLLMs with Unified General and Spatial Reasoning · signal_desk=talking · occurred_at=2025-07-11T00:00:00+00:00 · url=https://www.inclusion-ai.org/blog/m2-reasoning · raw={\"excerpt\":\"📖 Technical Report | 🤗 Hugging Face｜ 🤖 ModelScope\"}"},{"ref":"E60","kind":"event","title":"ABench: An Evolving Open-Source Benchmark","date":"2025-07-08T00:00:00+00:00","date_source":"rss.item_date","source_url":"https://www.inclusion-ai.org/blog/abench","signal_url":"https://onlylabs.fyi/signals/dd5ec616-9ce6-4065-8f85-a3688a2875b3","signal_json_url":"https://onlylabs.fyi/signals/dd5ec616-9ce6-4065-8f85-a3688a2875b3/signal.json","text":"post_published · ABench: An Evolving Open-Source Benchmark · signal_desk=talking · occurred_at=2025-07-08T00:00:00+00:00 · url=https://www.inclusion-ai.org/blog/abench · raw={\"excerpt\":\"GITHUB\"}"}]}