{"schema_version":"onlylabs.public_analysis_evidence.v1","title":"StepFun 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/stepfun","json_url":"https://onlylabs.fyi/analysis/stepfun/evidence.json","generated_at":"2026-06-11T15:10:51.228Z","org":{"slug":"stepfun","name":"StepFun","category":"neolab","category_label":"Neolab","dossier_url":"https://onlylabs.fyi/labs/stepfun"},"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":65,"web":0,"evidence":88,"signal_desks":{"hiring":0,"forks":2,"releases":31,"talking":0,"repos":27},"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":"stepfun-ai/Step-Realtime-CLI repository metadata","date":"2026-06-11T07:04:05.065184+00:00","date_source":null,"source_url":"https://github.com/stepfun-ai/Step-Realtime-CLI","signal_url":null,"signal_json_url":null,"text":"# stepfun-ai/Step-Realtime-CLI\n\nLanguage: TypeScript\n\nLicense: MIT\n\nStars: 18\n\nForks: 7\n\nOpen issues: 16\n\nCreated: 2026-06-01T06:00:45Z\n\nPushed: 2026-06-11T06:34:56Z\n\nDefault branch: main\n\nFork: no\n\nArchived: no\n\nREADME:\n# Step Realtime CLI\n\n<p align=\"center\">\n<a href=\"./README.md\">English</a> |\n<a href=\"./README_CN.md\">简体中文</a>\n</p>\n\n`step-realtime-cli` is a terminal-based AI coding assistant. You can interact with it via text or **realtime voice** for everyday tasks such as reading code, editing files, and running commands.\n\n## Demo\n\n![Step Realtime CLI demo](docs/assets/demo.gif)\n\n## Key capabilities\n\n- **Voice coding**: run `step voice` and, with headphones on, issue spoken instructions; the assistant parses repository context, applies edits, and confirms changes verbally.\n- **Text chat**: run `step` in any working directory to enter the interactive terminal UI and start a task with natural language.\n- **One-shot tasks**: submit a single request via `step exec \"...\"` and receive the result when execution completes.\n- **Session resumption**: session state is persisted automatically and can be resumed at any time via `step resume`.\n- **Read-only planning mode**: run `step exec --mode plan \"...\"` so the assistant only reads the code and proposes a plan, which the user reviews and approves before any changes are applied.\n\n## Quick start\n\n### Requirements\n\n- macOS / Linux, Node.js 20+\n- A StepFun API key (a single key may be used for both the coding model and realtime voice; a different provider's key may be configured for the coding side if preferred)\n\n### Choose your region\n\nStepFun operates two independent sites; pick the one that matches where your API key was issued. The two sites do **not** share accounts or keys.\n\n| Region | Console | API endpoint | Installer |\n| --- | --- | --- | --- |\n| Mainland China (default) | https://platform.stepfun.com/ | `https://api.stepfun.com` | `bash scripts/setup.sh` |\n| Overseas | https://platform.stepfun.ai/ | `https://api.stepfun.ai` | `bash scripts/setup-overseas.sh` |\n\n`scripts/setup-overseas.sh` runs the same flow as `scripts/setup.sh` and then rewrites `~/.step-cli/config.json` so both the realtime WebSocket and the mo"},{"ref":"P2","kind":"page","title":"stepfun-ai/Step-Realtime-CLI v0.1.0","date":"2026-06-11T07:04:04.589092+00:00","date_source":null,"source_url":"https://github.com/stepfun-ai/Step-Realtime-CLI/releases/tag/v0.1.0","signal_url":null,"signal_json_url":null,"text":"# step-realtime-cli v0.1.0\n\nRepository: stepfun-ai/Step-Realtime-CLI\n\nTag: v0.1.0\n\nPublished: 2026-06-11T06:32:05Z\n\nPrerelease: no\n\nRelease notes:\n## Step Realtime CLI v0.1.0 — Initial Release \n\n`step-realtime-cli` is a terminal-based AI coding assistant that you can drive via text or realtime voice for everyday tasks like reading code, editing files, and running commands. \n\n### Highlights \n- **Voice coding** — `step voice` lets you speak instructions; the assistant reads repository context, applies edits, and confirms changes verbally. \n- **Text chat** — `step` opens an interactive terminal UI for natural-language tasks in any working directory. \n- **One-shot tasks** — `step exec \"...\"` runs a single request and returns when it's done. \n- **Session resumption** — sessions persist automatically and can be resumed via `step resume`. \n- **Plan mode** — `step exec --mode plan \"...\"` lets the assistant read the code and propose a plan for review before any changes are applied."},{"ref":"P3","kind":"page","title":"stepfun-ai/Step-Video-T2V repository metadata","date":"2026-06-11T03:20:36.168882+00:00","date_source":null,"source_url":"https://github.com/stepfun-ai/Step-Video-T2V","signal_url":null,"signal_json_url":null,"text":"# stepfun-ai/Step-Video-T2V\n\nLanguage: Python\n\nLicense: MIT\n\nStars: 3185\n\nForks: 338\n\nOpen issues: 47\n\nCreated: 2025-02-08T08:46:51Z\n\nPushed: 2025-03-17T03:26:54Z\n\nDefault branch: main\n\nFork: no\n\nArchived: no\n\nREADME:\n<p align=\"center\">\n<img src=\"assets/logo.png\" height=100>\n</p>\n<div align=\"center\">\n<a href=\"https://yuewen.cn/videos\"><img src=\"https://img.shields.io/static/v1?label=Step-Video&message=Web&color=green\"></a> &ensp;\n<a href=\"https://arxiv.org/abs/2502.10248\"><img src=\"https://img.shields.io/static/v1?label=Tech Report&message=Arxiv&color=red\"></a> &ensp;\n<a href=\"https://x.com/StepFun_ai\"><img src=\"https://img.shields.io/static/v1?label=X.com&message=Web&color=blue\"></a> &ensp;\n</div>\n\n<div align=\"center\">\n<a href=\"https://huggingface.co/stepfun-ai/stepvideo-t2v\"><img src=\"https://img.shields.io/static/v1?label=Step-Video-T2V&message=HuggingFace&color=yellow\"></a> &ensp;\n<a href=\"https://huggingface.co/stepfun-ai/stepvideo-t2v-turbo\"><img src=\"https://img.shields.io/static/v1?label=Step-Video-T2V-Turbo&message=HuggingFace&color=yellow\"></a> &ensp;\n</div>\n\n## 🔥🔥🔥 News!!\n* Mar 17, 2025: 👋 We release the [Step-Video-TI2V](https://github.com/stepfun-ai/Step-Video-Ti2V), an image-to-video model based on Step-Video-T2V.\n* Feb 17, 2025: 👋 We release the inference code and model weights of Step-Video-T2V. [Download](https://huggingface.co/stepfun-ai/stepvideo-t2v)\n* Feb 17, 2025: 👋 We release the inference code and model weights of Step-Video-T2V-Turbo. [Download](https://huggingface.co/stepfun-ai/stepvideo-t2v-turbo)\n* Feb 17, 2025: 🎉 We have made our technical report available as open source. [Read](https://arxiv.org/abs/2502.10248)\n\n## Video Demos\n\n<table border=\"0\" style=\"width: 100%; text-align: center; margin-top: 1px;\">\n<tr>\n<td><video src=\"https://github.com/user-attachments/assets/9274b351-595d-41fb-aba3-f58e6e91603a\" width=\"100%\" controls autoplay loop muted></video></td>\n<td><video src=\"https://github.com/user-attachments/assets/2f6b3ad5-e93b-436b-98bc-4701182d8652\" width=\"100%\" controls autoplay loop muted></video></td>\n<td><video src=\"https://github.com/user-attachments/assets/67d20ee7-ad78-4b8f-80f6-3fdb00fb52d8\" width=\"100%\" controls "},{"ref":"P4","kind":"page","title":"stepfun-ai/Step-Video-TI2V repository metadata","date":"2026-06-11T03:20:35.532296+00:00","date_source":null,"source_url":"https://github.com/stepfun-ai/Step-Video-TI2V","signal_url":null,"signal_json_url":null,"text":"# stepfun-ai/Step-Video-TI2V\n\nLanguage: Python\n\nLicense: MIT\n\nStars: 375\n\nForks: 35\n\nOpen issues: 6\n\nCreated: 2025-03-06T04:42:11Z\n\nPushed: 2025-03-20T09:10:36Z\n\nDefault branch: main\n\nFork: no\n\nArchived: no\n\nREADME:\n<p align=\"center\">\n<img src=\"assets/logo.png\" height=100>\n</p>\n<div align=\"center\">\n<a href=\"https://yuewen.cn/videos\"><img src=\"https://img.shields.io/static/v1?label=Step-Video&message=Web&color=green\"></a> &ensp;\n<a href=\"https://arxiv.org/abs/2503.11251\"><img src=\"https://img.shields.io/static/v1?label=Tech Report&message=Arxiv&color=red\"></a> &ensp;\n<a href=\"https://x.com/StepFun_ai\"><img src=\"https://img.shields.io/static/v1?label=X.com&message=Web&color=blue\"></a> &ensp;\n</div>\n\n<div align=\"center\">\n<a href=\"https://huggingface.co/stepfun-ai/stepvideo-ti2v\"><img src=\"https://img.shields.io/static/v1?label=Step-Video-TI2V&message=HuggingFace&color=yellow\"></a> &ensp;\n</div>\n\n## 🔥🔥🔥 News!!\n* Mar 17, 2025: 👋 We release the inference code and model weights of Step-Video-TI2V. [Download](https://huggingface.co/stepfun-ai/stepvideo-ti2v)\n* Mar 17, 2025: 👋 We release a new TI2V benchmark [Step-Video-TI2V-Eval](https://github.com/stepfun-ai/Step-Video-TI2V/tree/main/benchmark/Step-Video-TI2V-Eval)\n* Mar 17, 2025: 👋 Step-Video-TI2V has been integrated into [ComfyUI-Stepvideo-ti2v](https://github.com/stepfun-ai/ComfyUI-StepVideo). Enjoy!\n* Mar 17, 2025: 🎉 We have made our technical report available as open source. [Read](https://arxiv.org/abs/2503.11251)\n\n## Motion Control\n\n<table border=\"0\" style=\"width: 100%; text-align: center; margin-top: 1px;\">\n<tr>\n<th style=\"width: 33%;\">战马跳跃</th>\n<th style=\"width: 33%;\">战马蹲下</th>\n<th style=\"width: 33%;\">战马向前奔跑，然后转身</th>\n</tr>\n<tr>\n<td><video src=\"https://github.com/user-attachments/assets/e664f45c-b8cd-4f89-9858-eaaef54aa0f6\" width=\"30%\" controls autoplay loop muted></video></td>\n<td><video src=\"https://github.com/user-attachments/assets/eb2d09b0-cc37-4f27-85c7-a31b6840fa69\" width=\"30%\" controls autoplay loop muted></video></td>\n<td><video src=\"https://github.com/user-attachments/assets/d17eba41-82f6-4ee2-8a99-3f21af112af0\" width=\"30%\" controls autoplay loop muted></video></td>\n</tr>\n</table>\n\n## Motion D"},{"ref":"P5","kind":"page","title":"stepfun-ai/Step-Audio repository metadata","date":"2026-06-11T03:20:35.523422+00:00","date_source":null,"source_url":"https://github.com/stepfun-ai/Step-Audio","signal_url":null,"signal_json_url":null,"text":"# stepfun-ai/Step-Audio\n\nLanguage: Python\n\nLicense: Apache-2.0\n\nStars: 27\n\nForks: 1\n\nOpen issues: 68\n\nCreated: 2025-02-11T05:35:12Z\n\nPushed: 2026-03-16T03:53:07Z\n\nDefault branch: main\n\nFork: no\n\nArchived: no\n\nREADME:\n<p align=\"left\">\n<a href=\"README_CN.md\">中文</a> &nbsp｜ &nbsp English&nbsp&nbsp ｜ &nbsp<a href=\"README_JP.md\">日本語</a>\n</p>\n<br><br>\n\n<img src=\"assets/QRCode.jpeg\" height=300>\n开发者微信交流群、Developer Group\n\n# This repository is no longer maintained, please refer to:\n## [Step-Audio2&Step-Audio2-mini](https://github.com/stepfun-ai/Step-Audio2) for End-to-end speech conversation\n## [Step-Audio-R1&Step-Audio-R1.1](https://github.com/stepfun-ai/Step-Audio-R1) for Speech Reasoning.\n## [Step-Audio-EditX](https://github.com/stepfun-ai/Step-Audio-EditX) for Audio Editing.\n\n# Step-Audio\n<p align=\"center\">\n<img src=\"assets/logo.png\" height=100>\n</p>\n<div align=\"center\">\n<a href=\"https://arxiv.org/abs/2502.11946\"><img src=\"https://img.shields.io/static/v1?label=Tech Report&message=Arxiv&color=red\"></a> &ensp;\n<a href=\"https://x.com/StepFun_ai\"><img src=\"https://img.shields.io/static/v1?label=X.com&message=Web&color=blue\"></a> &ensp;\n</div>\n\n<div align=\"center\">\n<a href=\"https://huggingface.co/stepfun-ai/Step-Audio-Chat\"><img src=\"https://img.shields.io/static/v1?label=Step-Audio-Chat&message=HuggingFace&color=yellow\"></a> &ensp;\n<a href=\"https://huggingface.co/stepfun-ai/Step-Audio-TTS-3B\"><img src=\"https://img.shields.io/static/v1?label=Step-Audio-TTS-3B&message=HuggingFace&color=yellow\"></a> &ensp;\n</div>\n<div align=\"center\">\n<a href=\"https://huggingface.co/stepfun-ai/Step-Audio-Tokenizer\"><img src=\"https://img.shields.io/static/v1?label=Step-Audio-Tokenier&message=HuggingFace&color=yellow\"></a> &ensp;\n<a href=\"https://huggingface.co/datasets/stepfun-ai/StepEval-Audio-360\"><img src=\"https://img.shields.io/static/v1?label=StepEval-Audio-360&message=HuggingFace&color=yellow\"></a> &ensp;\n</div>\n\n## 🔥🔥🔥 News!!\n* Aug 29, 2025: 👋 We release [Step-Audio 2 & Step-Audio 2 mini](https://github.com/stepfun-ai/Step-Audio2) and their corresponding inference [examples](examples.py). [Technical report](https://arxiv.org/pdf/2507.16632) is also updated.\n* Jun 10, 2025: 👋 We rel"},{"ref":"P6","kind":"page","title":"stepfun-ai/Step-Realtime-Console repository metadata","date":"2026-06-11T03:20:35.44204+00:00","date_source":null,"source_url":"https://github.com/stepfun-ai/Step-Realtime-Console","signal_url":null,"signal_json_url":null,"text":"# stepfun-ai/Step-Realtime-Console\n\nDescription: Step-Realtime-Console\n\nLanguage: JavaScript\n\nLicense: Apache-2.0\n\nStars: 74\n\nForks: 13\n\nOpen issues: 3\n\nCreated: 2025-03-12T10:09:47Z\n\nPushed: 2026-06-09T09:02:21Z\n\nDefault branch: main\n\nFork: no\n\nArchived: no\n\nREADME:\n# 阶跃星辰实时语音控制台 (Stepfun Realtime Console)\n\n中文 | [English](README-en.md)\n\n## 项目描述\n\n这是一个基于阶跃星辰实时语音 API 的前端演示项目，用于展示和测试阶跃星辰的实时语音对话功能。通过该项目，用户可以体验实时语音交互以及相关功能的调试与测试。项目提供了友好的用户界面，方便开发者和测试人员快速上手并了解阶跃星辰实时 API 的各项功能和特性。\n\n## 功能特点\n\n- **实时语音交互**：支持实时语音输入和输出，实现流畅的人机对话体验\n- **可视化音频波形**：使用 WaveSurfer.js 实现音频波形可视化\n- **自定义 AI 人设**：可以自定义 AI 的指令和人设，调整对话风格\n- **调试日志**：提供详细的调试日志，方便开发者追踪 API 调用和响应过程\n\n## 技术栈\n\n- **前端框架**：\n- Svelte 5 - 响应式前端框架\n- SvelteKit - 基于 Svelte 的服务端渲染 (SSR) 框架，类似于 Next.js 和 Nuxt.js，提供路由、服务端渲染和 API 端点等功能\n- **样式**：Tailwind CSS 4 - 实用优先的 CSS 框架\n- **UI 组件**：DaisyUI - 基于 Tailwind CSS 的组件库\n- **音频处理**：WaveSurfer.js - 音频可视化库\n- **构建工具**：Vite - 现代前端构建工具\n- **运行时**：Bun - 高性能 JavaScript 运行时\n- **语言**：TypeScript - 类型安全的 JavaScript 超集\n- **WebSocket**：Bun 原生 WebSocket - 用于实时通信\n\n## 安装与使用方法\n\n### 前置条件\n\n- 安装 Bun 运行时（必须，由于采用了 Bun 原生 WebSocket，因此不兼容 Node.js）\n\n### 安装步骤\n\n1. 安装 Bun 运行时：\n\n```bash\nnpm install -g bun\n```\n\n2. 克隆项目并进入项目目录：\n\n```bash\ngit clone https://github.com/stepfun-ai/Step-Realtime-Console\ncd Step-Realtime-Console\n```\n\n3. 安装项目依赖：\n\n```bash\nbun install\n```\n\n4. 运行开发服务器：\n\n```bash\nbun dev\n```\n\n项目将在 5173 端口运行（如被占用则顺序后延），同时 WebSocket 中转服务将在 8080 端口运行，请确保这些端口未被其他应用占用。请注意控制台输出的实际端口信息。\n\n5. 在浏览器中访问：\n```\nhttp://localhost:5173\n```\n\n### 构建生产版本\n\n```bash\nbun run build\n```\n\n构建后的文件将位于`build`目录中，您可通过`bun build/`来启动服务，注意最后的`/`不可省略，这条命令的完整版本其实是`bun build/index.js`。服务将在 3000 端口运行。\n\n如果您想自定义服务端口，可通过环境变量的方式，例如`PORT=3001 bun build/`，则服务会在 3001 端口运行。\n\n## 首次使用说明\n\n### 开发时首次加载\n\n开发时项目第一次页面加载时可能会比较慢，这是因为项目使用了 Lucide 图标库，该库在首次编译时需要较长时间进行处理，属于正常现象，请耐心等待。这种情况仅发生在开发阶段，生产版本不会有此问题。\n\n### 配置服务\n\n成功运行项目后，您需要在界面中点击**服务器设置**按钮并配置以下信息：\n\n1. **服务器地址**：wss://api.stepfun.com/v1/realtime\n\n2. **模型名称**：当前支持step-audio-2、step-audio-2-mini、step-audio-2-think、step-audio-2-mini-think 共四个版本模型\n\n3. **API Key**：您需要通过阶跃星辰开放平台获取 API 密钥。请访问 [阶跃星辰开放平台](https://platform.stepfun.com/) 注册并获取您的 API Key。\n\n4. **Voice**：音色设置（必填）。请填写您想要使用的音色值，例如：qingchunshaonv、wenrounansheng 等。\n\n填写完成后，"},{"ref":"P7","kind":"page","title":"stepfun-ai/ComfyUI-StepVideo repository metadata","date":"2026-06-11T03:20:35.378367+00:00","date_source":null,"source_url":"https://github.com/stepfun-ai/ComfyUI-StepVideo","signal_url":null,"signal_json_url":null,"text":"# stepfun-ai/ComfyUI-StepVideo\n\nLanguage: Python\n\nLicense: MIT\n\nStars: 43\n\nForks: 6\n\nOpen issues: 3\n\nCreated: 2025-03-11T04:12:38Z\n\nPushed: 2025-03-27T07:52:26Z\n\nDefault branch: main\n\nFork: no\n\nArchived: no\n\nREADME:\n# ComfyUI-StepVideo\nThis repository contains ComfyUI custom nodes for StepVideo.\n\n## TI2V\n### Workflow\n[TI2V.json](workflow/TI2V.json)\n![TI2V](workflow/TI2V.jpg)\n\n### Model Weights\nDownload the model weights from [this link](https://huggingface.co/stepfun-ai/stepvideo-ti2v).\n\n### Install\n1. Install [Step-Video-TI2V](https://github.com/stepfun-ai/Step-Video-TI2V)\n\n2. Install ComfyUI custom nodes for StepVideo\n```bash\ncd ComfyUI/custom_nodes\ngit clone https://github.com/stepfun-ai/ComfyUI-StepVideo.git \n```\n\n### Inference\n1. Launch Step-Video-TI2V remote_server\n```bash\ncd Step-Video-TI2V\npython api/call_remote_server.py --model_dir where_you_download_dir & ## We assume you have more than 4 GPUs available. This command will return the URL for both the caption API and the VAE API. Please use the returned URL as \"remote_server_url\" parameter in the \"TI2V\" node.\n```\n\n2. Launch ComfyUI\n```bash\ncd ComfyUI\npython main.py\n```\n\n## TI2V_API\n### Workflow\n[TI2V_API.json](workflow/TI2V_API.json)\n![TI2V_API](workflow/TI2V_API.jpg)\n\n### Usage\napi_url: https://api.stepfun.com/v1/video/generations\n\napi_key: get api_key from https://platform.stepfun.com\n\n## Todo\n- [x] TI2V node\n- [x] TI2V_API node\n- [ ] T2V node\n- [ ] T2V_API node"},{"ref":"P8","kind":"page","title":"stepfun-ai/Step1X-Edit repository metadata","date":"2026-06-11T03:20:35.288048+00:00","date_source":null,"source_url":"https://github.com/stepfun-ai/Step1X-Edit","signal_url":null,"signal_json_url":null,"text":"# stepfun-ai/Step1X-Edit\n\nDescription: A SOTA open-source image editing model, which aims to provide comparable performance against the closed-source models like GPT-4o and Gemini 2 Flash.\n\nLanguage: Python\n\nLicense: Apache-2.0\n\nStars: 2224\n\nForks: 102\n\nOpen issues: 32\n\nCreated: 2025-04-23T09:53:08Z\n\nPushed: 2026-04-29T14:31:44Z\n\nDefault branch: main\n\nFork: no\n\nArchived: no\n\nREADME:\n<div align=\"center\">\n<img src=\"assets/logo.png\" height=100>\n</div>\n<div align=\"center\">\n<a href=\"https://step1x-edit.github.io/\"><img src=\"https://img.shields.io/static/v1?label=Project%20Page&message=Web&color=green\"></a> &ensp;\n<a href=\"https://arxiv.org/abs/2504.17761\"><img src=\"https://img.shields.io/static/v1?label=Step1X-Edit&message=Arxiv&color=red\"></a> &ensp;\n<a href=\"https://arxiv.org/abs/2511.22625\"><img src=\"https://img.shields.io/static/v1?label=ReasonEdit&message=Arxiv&color=red\"></a> &ensp;\n<a href=\"assets/WeChat.jpg\">\n<img src=\"https://img.shields.io/static/v1?label=WeChat&message=Add%20Me&color=green&logo=wechat&logoColor=white\">\n</a>\n\n<a href=\"https://huggingface.co/stepfun-ai/Step1X-Edit\"><img src=\"https://img.shields.io/static/v1?label=Model&message=HuggingFace&color=yellow\"></a> &ensp;\n<a href=\"https://huggingface.co/datasets/stepfun-ai/GEdit-Bench\"><img src=\"https://img.shields.io/static/v1?label=GEdit-Bench&message=HuggingFace&color=yellow\"></a> &ensp;\n[![Run on Replicate](https://replicate.com/zsxkib/step1x-edit/badge)](https://replicate.com/zsxkib/step1x-edit) &ensp;\n<a href=\"https://discord.gg/j3qzuAyn\"><img src=\"https://img.shields.io/static/v1?label=Discord%20Channel&message=Discord&color=purple\"></a> &ensp;\n</div>\n\n## 🔥🔥🔥 News!!\n* Apr 29, 2026: 🎉 Step Image Edit 2 is now live — a lightweight model designed for ultra-fast response and high-quality output, delivering a real-time interactive creation experience. It can complete image generation and editing tasks within 2 seconds. Feel free to try it out and share your feedback ✨✨✨\n\nTry it here (StepFun Open Platform): [https://platform.stepfun.com/docs/zh/guides/models/step-image-edit-2](https://platform.stepfun.com/docs/zh/guides/models/step-image-edit-2)\n\nAPI documentation: [https://platform.stepfun.co"},{"ref":"P9","kind":"page","title":"stepfun-ai/Step-Realtime-Python-Demo repository metadata","date":"2026-06-11T03:20:34.753324+00:00","date_source":null,"source_url":"https://github.com/stepfun-ai/Step-Realtime-Python-Demo","signal_url":null,"signal_json_url":null,"text":"# stepfun-ai/Step-Realtime-Python-Demo\n\nDescription: Step Realtime Demo\n\nLanguage: Python\n\nStars: 3\n\nForks: 1\n\nOpen issues: 0\n\nCreated: 2025-04-30T06:19:24Z\n\nPushed: 2025-05-06T09:22:02Z\n\nDefault branch: main\n\nFork: no\n\nArchived: no\n\nREADME:\n# VAD模式演示\n\n中文 | [English](README-en.md)\n\n**重要提示**：\n此演示不支持声学回声消除（AEC）。请使用耳机！\n\n## 如何运行\n\n### 初始化运行环境\n\n1. 按照[uv安装指南](https://docs.astral.sh/uv/getting-started/installation/)安装uv\n2. 按照[PyAudio文档](https://pypi.org/project/PyAudio/)安装pyaudio\n\n### 更新配置\n\n将API_KEY替换为您自己的密钥\n\n### 运行Python脚本\n\n```shell\nuv run vad_mode.py\n```\n\n如果您遇到 \"ImportError: python-socks is required to use a SOCKS proxy\"，请停止您的代理"},{"ref":"P10","kind":"page","title":"stepfun-ai/Step1X-3D repository metadata","date":"2026-06-11T03:20:34.639353+00:00","date_source":null,"source_url":"https://github.com/stepfun-ai/Step1X-3D","signal_url":null,"signal_json_url":null,"text":"# stepfun-ai/Step1X-3D\n\nDescription: Step1X-3D: Towards High-Fidelity and Controllable Generation of Textured 3D Assets\n\nLanguage: Python\n\nLicense: Apache-2.0\n\nStars: 869\n\nForks: 59\n\nOpen issues: 40\n\nCreated: 2025-05-13T03:42:54Z\n\nPushed: 2025-09-08T22:32:12Z\n\nDefault branch: main\n\nFork: no\n\nArchived: no\n\nREADME:\n<p align=\"left\">\n<a href=\"README_CN.md\">中文</a> &nbsp｜ &nbsp English&nbsp&nbsp \n</p>\n\n<h1 align=\"center\"> Step1X-3D: Towards High-Fidelity and Controllable<br>Generation of Textured 3D Assets</h1>\n\n<div align=\"center\">\n<a href=https://huggingface.co/spaces/stepfun-ai/Step1X-3D target=\"_blank\"><img src=https://img.shields.io/static/v1?label=Online%20Demo&message=HuggingFace&color=yellow></a>\n<a href=https://huggingface.co/stepfun-ai/Step1X-3D target=\"_blank\"><img src=https://img.shields.io/static/v1?label=Model&message=HuggingFace&color=yellow></a>\n<a href=https://arxiv.org/abs/2505.07747 target=\"_blank\"><img src=https://img.shields.io/static/v1?label=Tech%20Report&message=Arxiv&color=red></a>\n<a href=https://stepfun-ai.github.io/Step1X-3D/ target=\"_blank\"><img src= https://img.shields.io/static/v1?label=Project%20Page&message=Web&color=green></a>\n</div>\n\n<p align=\"center\">\n<img src=\"assets/stepfun_illusions_logo.jpeg\" width=\"100%\">\n</p>\n\n<div align=\"center\">\n<img src=\"./assets/step1x-3d-teaser.png\" width=\"100%\">\n</div>\n\n<div align=\"left\">\n<p><b>Step1X-3D demonstrates the capability to generate 3D assets with high-fidelity geometry and versatile texture maps, while maintaining exceptional alignment between surface geometry and texture mapping. From left to right, we sequentially present: the base geometry (untextured), followed by cartoon-style, sketch-style, and photorealistic 3D asset generation results.</b></p>\n</div>\n\n## 🔥🔥🔥 Latest News!!\n* June 26, 2025: 👋 We release the data preprocessing for shape VAE and diffusion, including advanced watertight method using depth_test and winding_number proposed by [CraftsMan3D](https://github.com/wyysf-98/CraftsMan3D) in path \"Step1X-3D/data/watertight_and_sampling.py\"! \n* June 9, 2025: 👋 We release the multi-views render code for texture generation model training in path \"Step1X-3D/data/ig2mv/render\"!\n* May"},{"ref":"P11","kind":"page","title":"stepfun-ai/InfiniteHBD-Trace repository metadata","date":"2026-06-11T03:20:34.632774+00:00","date_source":null,"source_url":"https://github.com/stepfun-ai/InfiniteHBD-Trace","signal_url":null,"signal_json_url":null,"text":"# stepfun-ai/InfiniteHBD-Trace\n\nLicense: Apache-2.0\n\nStars: 18\n\nForks: 1\n\nOpen issues: 0\n\nCreated: 2025-05-21T02:16:57Z\n\nPushed: 2025-05-21T02:21:04Z\n\nDefault branch: main\n\nFork: no\n\nArchived: no\n\nREADME:\n# InfiniteHBD-Trace\n\n> 🏆 This work has been **accepted to [ACM SIGCOMM 2025](https://conferences.sigcomm.org/sigcomm/2025/)**. It provides a real-world fault trace dataset for large-scale GPU clusters used in LLM pretraining workloads.\n\nThis project open-sources fault trace data from **400** GPU servers, with fault events impacting up to **231 distinct servers**. These servers are randomly picked across a GPU cluster comprising hundreds of nodes and thousands of GPUs. The dataset, covering **348 days of fault data starting from March 30, 2024**, is primarily used to support large-scale **pretraining of large language models (LLMs)** and provides realistic fault-tolerance testing data for simulation experiments. The dataset has been used in the *InfiniteHBD*, which includes detailed statistical analysis (Paper link: [https://arxiv.org/abs/2502.03885](https://arxiv.org/abs/2502.03885)).\n\n## Project Description\n\nThis repository contains the following files:\n- **fault_trace.json**: The fault trace dataset.\n- **fault_statistics.json**: Hierarchical breakdown of fault types including Level, Class, and Description, with raw counts.\n- **README.md**: Project documentation.\n\n## Trace Description\n\nThe trace records a series of node fault events in the GPU cluster. Each event is stored in JSON format. An example is shown below:\n\n```json\n{\n\"node_id\": \"067eb1e2-ea0b-4069-b64e-5df892642f88\",\n\"event_time\": 8.6112,\n\"event_type\": \"fault_start\",\n\"fault_type\": {\n\"Level\": \"Hardware Failure\",\n\"Class\": \"GPU\",\n\"Desc\": \"GPU xid Error\"\n}\n},\n{\n\"node_id\": \"067eb1e2-ea0b-4069-b64e-5df892642f88\",\n\"event_time\": 8.8896,\n\"event_type\": \"fault_end\",\n\"fault_type\": {\n\"Level\": \"Hardware Failure\",\n\"Class\": \"GPU\",\n\"Desc\": \"GPU xid Error\"\n}\n}\n```\n\n### Field Descriptions\n\n- **node_id** \nUUID as node identifier. The specific information about the nodes has been anonymized.\n\n- **event_time** \nThe relative time of the event (unit: days). The time value is calculated relative to the first event in the tra"},{"ref":"P12","kind":"page","title":"stepfun-ai/Step-Audio2 repository metadata","date":"2026-06-11T03:20:34.626901+00:00","date_source":null,"source_url":"https://github.com/stepfun-ai/Step-Audio2","signal_url":null,"signal_json_url":null,"text":"# stepfun-ai/Step-Audio2\n\nDescription: Step-Audio 2 is an end-to-end multi-modal large language model designed for industry-strength audio understanding and speech conversation.\n\nLanguage: Python\n\nLicense: Apache-2.0\n\nStars: 1460\n\nForks: 107\n\nOpen issues: 55\n\nCreated: 2025-07-15T09:14:32Z\n\nPushed: 2026-03-16T04:06:21Z\n\nDefault branch: main\n\nFork: no\n\nArchived: no\n\nREADME:\n# Step-Audio 2\n\n<div align=\"center\">\n<img src=\"assets/logo.png\" height=100>\n</div>\n\n<div align=\"center\" style=\"line-height: 1;\">\n<a href=\"https://stepfun.com/\" target=\"_blank\"><img alt=\"Homepage\" src=\"https://img.shields.io/badge/Homepage-StepFun-white?logo=StepFun&logoColor=white\"/></a> &ensp;\n<a href=\"https://x.com/StepFun_ai\" target=\"_blank\"><img alt=\"Twitter Follow\" src=\"https://img.shields.io/badge/Twitter-StepFun-white?logo=x&logoColor=white\"/></a> &ensp;\n<a href=\"https://discord.com/invite/XHheP5Fn\" target=\"_blank\"><img alt=\"Discord\" src=\"https://img.shields.io/badge/Discord-StepFun-white?logo=discord&logoColor=white\"/></a>\n</div>\n<div align=\"center\">\n<a href=\"https://huggingface.co/stepfun-ai/Step-Audio-2-mini\"><img src=\"https://img.shields.io/static/v1?label=Step-Audio-2-mini&message=HuggingFace&color=yellow\"></a> &ensp;\n<a href=\"https://huggingface.co/stepfun-ai/Step-Audio-2-mini-Base\"><img src=\"https://img.shields.io/static/v1?label=Step-Audio-2-mini-Base&message=HuggingFace&color=yellow\"></a> &ensp;\n<a href=\"https://huggingface.co/stepfun-ai/Step-Audio-2-mini-Think\"><img src=\"https://img.shields.io/static/v1?label=Step-Audio-2-mini-Think&message=HuggingFace&color=yellow\"></a>\n</div>\n<div align=\"center\">\n<a href=\"https://arxiv.org/abs/2507.16632\"><img src=\"assets/arxiv.svg\"></a> &ensp;\n<a href=\"https://github.com/stepfun-ai/Step-Audio2/blob/main/LICENSE\"><img alt=\"License\" src=\"https://img.shields.io/badge/License-Apache%202.0-blue?&color=blue\"/></a>\n</div>\n\n## 🔥🔥🔥 News!!\n<!-- * Aug 29, 2025: 👋 We release the # TODO -->\n* Sep 15, 2025: 👋 We release [Step-Audio 2 mini Think](https://huggingface.co/stepfun-ai/Step-Audio-2-mini-Think) and its corresponding [examples](examples-think.py).\n* Sep 3, 2025: 👋 We release our [vLLM backend](https://github.com/stepfun-ai/vllm/tree/step-aud"},{"ref":"P13","kind":"page","title":"stepfun-ai/StepMesh repository metadata","date":"2026-06-11T03:20:34.281081+00:00","date_source":null,"source_url":"https://github.com/stepfun-ai/StepMesh","signal_url":null,"signal_json_url":null,"text":"# stepfun-ai/StepMesh\n\nLanguage: C++\n\nLicense: Apache-2.0\n\nStars: 367\n\nForks: 41\n\nOpen issues: 11\n\nCreated: 2025-07-21T11:47:08Z\n\nPushed: 2026-01-28T11:26:52Z\n\nDefault branch: main\n\nFork: no\n\nArchived: no\n\nREADME:\n# StepMesh: A High-Performance, Low-Latency Communication Library for Attention-FFN Disaggregation\n\nStepMesh is a communication library designed to provide high-performance and low-latency\ncommunication for Attention-FFN decoupling architectures.\n\nThe codebase is developed upon [BytePS](https://github.com/bytedance/ps-lite).\n\n## Overview\n\nThe following diagram illustrates the architecture of the **StepMesh** communication framework with the follow highlights.\n\n- **Unified and Flexible ZeroCopy PushPull Interface**: StepMesh provides a single, efficient interface (*ZBatchPushPull*) for tensor transfer.\n*ZBatchPushPull* interface allows users to tailor communication strategies to specific application needs.\n- **Communication Across Multiple Accelerators**: Designed to work with different computing Accelerators,\nStepMesh leverages RDMA to achieve high-speed, low-latency communication\nand is architected to easily incorporate support for new computing chips.\n- **Low Latency**: StepMesh is engineered to provide the ultra-low latency communication essential for Attention-FFN decoupling architectures.\nIts performance is on par with the limits imposed by hardware components.\n\n<!-- ![StepMesh Framework](./docs/images/framework.png) -->\n\n- **AFTensorWorker API**: This API is responsible for managing the communication operations performed by attention nodes. It provides methods such as `Wait` and `PushPull` for synchronization and data transfer.\n- **AFTensorServer API**: This API handles the communication operations performed by FFN nodes. It includes methods like `GetBatch` and `Respond` for retrieving and responding to requests from workers.\n\n#### StepMesh Core\n\n- **NetSend/NetRecv Threads**: These threads are part of the **StepMesh Core** and are responsible for handling network send and receive operations. They interact with various backends to manage data transfers efficiently.\n\n#### Backends\n\n- **RDMATransport**: This backend is responsible for managing **Re"},{"ref":"P14","kind":"page","title":"stepfun-ai/Step3 repository metadata","date":"2026-06-11T03:20:33.550173+00:00","date_source":null,"source_url":"https://github.com/stepfun-ai/Step3","signal_url":null,"signal_json_url":null,"text":"# stepfun-ai/Step3\n\nLicense: Apache-2.0\n\nStars: 453\n\nForks: 12\n\nOpen issues: 5\n\nCreated: 2025-07-25T03:11:20Z\n\nPushed: 2025-08-10T15:26:18Z\n\nDefault branch: main\n\nFork: no\n\nArchived: no\n\nREADME:\n<div align=\"center\">\n<picture>\n<img src=\"figures/stepfun-logo.png\" width=\"30%\" alt=\"StepFun: Cost-Effective Multimodal Intelligence\">\n</picture>\n</div>\n\n<hr>\n\n<div align=\"center\" style=\"line-height:1\">\n<a href=\"https://stepfun.com/\" target=\"_blank\"><img alt=\"Chat\" src=\"https://img.shields.io/badge/Chat-StepFun-ff6b6b?color=1783ff&logoColor=white\"/></a>\n<a href=\"https://stepfun.com/\" target=\"_blank\"><img alt=\"Homepage\" src=\"https://img.shields.io/badge/Homepage-StepFun-white?logo=StepFun&logoColor=white\"/></a>\n</div>\n\n<div align=\"center\" style=\"line-height: 1;\">\n<a href=\"https://huggingface.co/collections/stepfun-ai/step3-688a3d652dbb45d868f9d42d\" target=\"_blank\"><img alt=\"Hugging Face\" src=\"https://img.shields.io/badge/%F0%9F%A4%97%20Hugging%20Face-StepFun-ffc107?color=ffc107&logoColor=white\"/></a>\n<a href=\"https://www.modelscope.cn/models/stepfun-ai/step3\" target=\"_blank\"><img alt=\"ModelScope\" src=\"https://img.shields.io/badge/ModelScope-StepFun-white?logo=modelscope&logoColor=white\"/></a>\n<a href=\"https://x.com/StepFun_ai\" target=\"_blank\"><img alt=\"Twitter Follow\" src=\"https://img.shields.io/badge/Twitter-StepFun-white?logo=x&logoColor=white\"/></a>\n</div>\n\n<div align=\"center\" style=\"line-height: 1;\">\n<a href=\"https://discord.com/invite/XHheP5Fn\" target=\"_blank\"><img alt=\"Discord\" src=\"https://img.shields.io/badge/Discord-StepFun-white?logo=discord&logoColor=white\"/></a>\n<a href=\"LICENSE\"><img alt=\"License\" src=\"https://img.shields.io/badge/License-Apache%202.0-blue?&color=blue\"/></a>\n</div>\n\n<div align=\"center\">\n<b>📰&nbsp;&nbsp;<a href=\"https://stepfun.ai/research/step3\">Step3 Model Blog</a></b> &nbsp;&nbsp;&nbsp; | &nbsp;&nbsp;&nbsp; <b>📄&nbsp;&nbsp;<a href=\"https://arxiv.org/abs/2507.19427\">Step3 System Tech Report</a></b>\n</div>\n\n## Introduction\n\nStep3 is our cutting-edge multimodal reasoning model—built on a Mixture-of-Experts architecture with 321B total parameters and 38B active. \nIt is designed end-to-end to minimize decoding costs while delivering top-tier pe"},{"ref":"P15","kind":"page","title":"stepfun-ai/NextStep-1 repository metadata","date":"2026-06-11T03:20:33.432599+00:00","date_source":null,"source_url":"https://github.com/stepfun-ai/NextStep-1","signal_url":null,"signal_json_url":null,"text":"# stepfun-ai/NextStep-1\n\nDescription: [🚀 ICLR 2026 Oral] NextStep-1: SOTA Autogressive Image Generation with Continuous Tokens. A research project developed by the StepFun’s Multimodal Intelligence team.\n\nLanguage: Python\n\nLicense: Apache-2.0\n\nStars: 689\n\nForks: 26\n\nOpen issues: 0\n\nCreated: 2025-08-14T08:50:25Z\n\nPushed: 2026-02-27T17:05:44Z\n\nDefault branch: main\n\nFork: no\n\nArchived: no\n\nREADME:\n# NextStep-1: Toward Autoregressive Image Generation with Continuous Tokens at Scale\n\n<div align=\"center\">\n\n[![Homepage](https://img.shields.io/static/v1?label=Homepage&message=Project%20Page&color=blue&logo=home)](https://stepfun.ai/research/en/nextstep1)&nbsp;[![huggingface weights](https://img.shields.io/badge/%F0%9F%A4%97%20Weights-StepFun/NextStep1-yellow)](https://huggingface.co/collections/stepfun-ai/nextstep-1-689d80238a01322b93b8a3dc)&nbsp;[![arXiv:2508.10711](https://img.shields.io/badge/arXiv-2508.10711-b31b1b.svg)](https://arxiv.org/abs/2508.10711)&nbsp;[![Blog](https://img.shields.io/badge/Blog-NextStep1-blue)](https://stepfun-ai.github.io/NextStep-1/nextstep_1_blog/)&nbsp;[![Blog](https://img.shields.io/badge/Blog-NextStep1.1-blue)](https://stepfun-ai.github.io/NextStep-1/nextstep_1p1_blog/)\n\n</div>\n\n> Autoregressive models—generating content step-by-step like reading a sentence—excel in language but struggle with images. Traditionally, they either depend on costly diffusion models or compress images into discrete, lossy tokens via vector quantization (VQ).\n>\n> NextStep-1 takes a different path: a 14B-parameter autoregressive model that works directly with continuous image tokens, preserving the full richness of visual data. It models sequences of discrete text tokens and continuous image tokens jointly—using a standard LM head for text and a lightweight 157M-parameter flow matching head for visuals. This unified next-token prediction framework is simple, scalable, and capable of producing stunningly detailed images.\n\n<div align=\"center\">\n<img width=\"720\" alt=\"t2i_demo\" src=\"./assets/t2i_demo.gif\">\n</div>\n\n<div align=\"center\">\n<img width=\"720\" alt=\"edit_demo\" src=\"./assets/edit_demo.gif\">\n</div>\n\n## 🔥 News\n\n- **Feb. 25, 2026**: **vLLM-Omni** supports high "},{"ref":"P16","kind":"page","title":"stepfun-ai/StepFun-Prover-Preview repository metadata","date":"2026-06-11T03:20:33.395243+00:00","date_source":null,"source_url":"https://github.com/stepfun-ai/StepFun-Prover-Preview","signal_url":null,"signal_json_url":null,"text":"# stepfun-ai/StepFun-Prover-Preview\n\nDescription: Large language models designed for formal theorem proving through tool-integrated reasoning.\n\nLicense: Apache-2.0\n\nStars: 35\n\nForks: 1\n\nOpen issues: 0\n\nCreated: 2025-08-13T09:19:23Z\n\nPushed: 2025-08-13T10:39:32Z\n\nDefault branch: main\n\nFork: no\n\nArchived: no\n\nREADME:\n<p align=\"center\">\n<img src=\"assets/logo.png\" width=\"250px\"><br>\n</p>\n\n<div align=\"center\"> \n<a href=\"https://arxiv.org/abs/2507.20199\"><img src=\"https://img.shields.io/static/v1?label=Technical%20Report&message=Arxiv&color=red\"></a> &ensp;\n<a href=\"https://huggingface.co/stepfun-ai/StepFun-Prover-Preview-32B\"><img src=\"https://img.shields.io/static/v1?label=Model&message=HuggingFace&color=yellow\"></a> &ensp;\n</div>\n<br>\n\n## Introduction\n\nWe introduce StepFun-Prover-Preview, large language models designed for formal theorem proving through tool-integrated reasoning. Using a reinforcement learning pipeline that incorporates tool-based interactions, StepFun-Prover Preview can achieve strong performance in generating Lean 4 proofs with minimal sampling. Our approach enables the model to emulate human-like problem-solving strategies by iteratively refining proofs based on real-time environment feedback. On the miniF2F-test benchmark, StepFun-Prover-Preview-32B achieves a pass@1 success rate of 70%. Please refer to our [technical report](https://arxiv.org/abs/2507.20199) for more details.\n\n<p align=\"center\">\n<img width=\"60%\" src=\"assets/fig1.png\">\n</p>\n\n**Figure 1:** Performance comparison on MiniF2F-test. y-axis shows the pass@1, which is computed by averaging 32 trials; while x-axis denotes the maximum number the provers are allowed to interact with Lean4-REPL before getting successful proof. Note both DeepSeek-Prover and Kimina-Prover utilize at least 32K token context length. Stepfun-Prover was evaluated using 20K context window including feedback from Lean4-REPL.\n\n## Methodology\n\n<p align=\"center\">\n<img width=\"60%\" src=\"assets/fig2.png\">\n</p>\n\n**Figure 2:** Our training pipeline. Left. Tool-integrated RL and iterative RL-SFT cycle. The upper left illustrates data preparation. Right. Tool-integrated reasoning pattern.\n\n## Experimental Results\n\n<div ali"},{"ref":"P17","kind":"page","title":"stepfun-ai/StepFun-Formalizer repository metadata","date":"2026-06-11T03:20:32.822689+00:00","date_source":null,"source_url":"https://github.com/stepfun-ai/StepFun-Formalizer","signal_url":null,"signal_json_url":null,"text":"# stepfun-ai/StepFun-Formalizer\n\nDescription: StepFun-Formalizer: Unlocking the Autoformalization Potential of LLMs through Knowledge-Reasoning Fusion\n\nLanguage: Python\n\nLicense: Apache-2.0\n\nStars: 30\n\nForks: 3\n\nOpen issues: 0\n\nCreated: 2025-08-19T03:34:06Z\n\nPushed: 2025-08-19T03:34:42Z\n\nDefault branch: master\n\nFork: no\n\nArchived: no\n\nREADME:\n# StepFun-Formalizer: Unlocking the Autoformalization Potential of LLMs through Knowledge-Reasoning Fusion\n\n<p align=\"center\">\n<img src=\"assets/logo.png\" width=\"250px\"><br>\n</p>\n\n<div align=\"center\"> \n<a href=\"https://www.arxiv.org/abs/2508.04440\"><img src=\"https://img.shields.io/static/v1?label=Paper&message=Arxiv&color=red\"></a> &ensp;\n<a href=\"https://huggingface.co/stepfun-ai/StepFun-Formalizer-32B\"><img src=\"https://img.shields.io/static/v1?label=Model&message=HuggingFace&color=yellow\"></a> &ensp;\n<a href=\"https://github.com/stepfun-ai/StepFun-Formalizer\"><img src=\"https://img.shields.io/static/v1?label=Code&message=Github&color=blue\"></a> &ensp;\n</div>\n<br>\n\n## Introduction\n\nWe introduce StepFun-Formalizer, a family of large language models designed to translate natural-language mathematical problems into formal statements in Lean 4. Through the fusion of formal knowledge and informal-to-formal reasoning capability, StepFun-Formalizer achieves strong performance on autoformalization tasks. Evaluated with [BEq](https://github.com/Purewhite2019/rethinking_autoformalization) verification on mainstream benchmarks including [FormalMATH-Lite](https://huggingface.co/datasets/SphereLab/FormalMATH-Lite), [ProverBench](https://huggingface.co/datasets/deepseek-ai/DeepSeek-ProverBench), and [CombiBench](https://huggingface.co/datasets/AI-MO/CombiBench), StepFun-Formalizer matches or exceeds all prior general-purpose and specialized autoformalization models of comparable scale. Please refer to our [paper](https://arxiv.org/abs/2508.04440) for more details.\n\n<p align=\"center\">\n<img width=\"80%\" src=\"assets/fig1.png\">\n</p>\n\n**Figure 1: A case study to demonstrate the impact of formal knowledge and informal-to-formal reasoning capability on autoformalization models.** It shows that general-purpose models without formal knowledge make "},{"ref":"P18","kind":"page","title":"stepfun-ai/Step-Audio-EditX repository metadata","date":"2026-06-11T03:20:32.697679+00:00","date_source":null,"source_url":"https://github.com/stepfun-ai/Step-Audio-EditX","signal_url":null,"signal_json_url":null,"text":"# stepfun-ai/Step-Audio-EditX\n\nDescription: A powerful 3B-parameter, LLM-based Reinforcement Learning audio edit model excels at editing emotion, speaking style, and paralinguistics, and features robust zero-shot text-to-speech\n\nLanguage: Python\n\nLicense: Apache-2.0\n\nStars: 929\n\nForks: 69\n\nOpen issues: 37\n\nCreated: 2025-10-29T11:54:17Z\n\nPushed: 2026-04-09T02:27:46Z\n\nDefault branch: main\n\nFork: no\n\nArchived: no\n\nREADME:\n# Step-Audio-EditX\n<p align=\"center\">\n<img src=\"assets/logo.png\" height=100>\n</p>\n\n<div align=\"center\">\n<a href=\"https://stepaudiollm.github.io/step-audio-editx/\"><img src=\"https://img.shields.io/static/v1?label=Demo%20Page&message=Web&color=green\"></a> &ensp;\n<a href=\"https://arxiv.org/abs/2511.03601\"><img src=\"https://img.shields.io/static/v1?label=Tech%20Report&message=Arxiv&color=red\"></a> &ensp;\n<a href=\"https://huggingface.co/stepfun-ai/Step-Audio-EditX\"><img src=\"https://img.shields.io/static/v1?label=Step-Audio-EditX&message=HuggingFace&color=yellow\"></a> &ensp;\n\n<a href=\"https://modelscope.cn/models/stepfun-ai/Step-Audio-EditX\"><img src=\"https://img.shields.io/static/v1?label=Step-Audio-EditX&message=ModelScope&color=blue\"></a> &ensp;\n<a href=\"https://huggingface.co/spaces/stepfun-ai/Step-Audio-EditX\"><img src=\"https://img.shields.io/static/v1?label=Space%20Playground&message=HuggingFace&color=yellow\"></a> &ensp;\n<a href=\"https://www.stepfun.com/studio/audio?tab=edit\"><img src=\"https://img.shields.io/static/v1?label=Audio%20Studio&message=StepFun&color=blue\"></a> &ensp;\n</div>\n\n## 🔥🔥🔥 News!!！\n* Jan 29, 2026: \n* 🧩 New Model Release: \n* Better performance, with an overall improvement of over 4%.\n* More **paralinguistic** tags have been added, including **`exhale`**, **`snort`**, **`inhale`**, **`chuckle`**, **`clears throat`**, **`giggle`**.\n* Welcome to try out at [StepFun Audio Studio](https://www.stepfun.com/studio/audio?tab=edit)\n* 💻 We release the **SFT**, **DPO** and **GRPO** training code.\n* 🌟 Training and inference for **vLLM** are now supported. Thanks to the vLLM team!\n* Nov 28, 2025: 🚀 New Model Release: Now supporting **`Japanese`** and **`Korean`** languages.\n* Nov 23, 2025: 📊 [Step-Audio-Edit-Benchmark](https://github."},{"ref":"P19","kind":"page","title":"stepfun-ai/Step-Audio-R1 repository metadata","date":"2026-06-11T03:20:32.442397+00:00","date_source":null,"source_url":"https://github.com/stepfun-ai/Step-Audio-R1","signal_url":null,"signal_json_url":null,"text":"# stepfun-ai/Step-Audio-R1\n\nLanguage: Python\n\nLicense: Apache-2.0\n\nStars: 673\n\nForks: 48\n\nOpen issues: 20\n\nCreated: 2025-11-11T02:38:13Z\n\nPushed: 2026-04-29T04:15:21Z\n\nDefault branch: main\n\nFork: no\n\nArchived: no\n\nREADME:\n# Step-Audio-R1/R1.5\n<p align=\"center\">\n<img src=\"assets/logo.png\" height=100>\n</p>\n\n<div align=\"center\">\n<a href=\"https://modelscope.cn/studios/stepfun-ai/Step-Audio-R1\"><img src=\"https://img.shields.io/static/v1?label=Demo%20Page&message=Web&color=green\"></a> &ensp;\n<a href=\"https://arxiv.org/abs/2604.25719\"><img src=\"https://img.shields.io/static/v1?label=Step-Audio-R1.5&message=Arxiv&color=red\"></a> &ensp;\n<a href=\"https://huggingface.co/stepfun-ai/Step-Audio-R1\"><img src=\"https://img.shields.io/static/v1?label=Step-Audio-R1.1&message=HuggingFace&color=yellow\"></a> &ensp;\n<a href=\"https://modelscope.cn/models/stepfun-ai/Step-Audio-R1.1\"><img src=\"https://img.shields.io/static/v1?label=Step-Audio-R1.1&message=ModelScope&color=blue\"></a> &ensp;\n<a href=\"https://www.stepfun.com/studio/audio?tab=conversation\"><img src=\"https://img.shields.io/static/v1?label=Space%20Playground&message=Studio&color=yellow\"></a> &ensp;\n</div>\n\n## 🔥🔥🔥 News!!\n* Apr 29, 2026: 🚀 We release the technical report of **Step-Audio-R1.5** ([ArXiv](https://arxiv.org/abs/2604.25719); [PDF](Step-Audio-R1.5.pdf)).\n* Apr 29, 2026: 📦 We open-source three in-house benchmarks from **Step-Audio-R1.5** under `benchmarks/Step-Audio-R1.5/`: `step_caption`, `step_spqa`, and `step_dialogue_understanding`.\n* Jan 14, 2026: 🚀 We release the inference code and model weights of **Step-Audio-R1.1** ([HuggingFace](https://huggingface.co/stepfun-ai/Step-Audio-R1.1); [ModelScope](https://modelscope.cn/models/stepfun-ai/Step-Audio-R1.1))\n* Nov 27, 2025: 🎉 We release the inference code and model weights of **Step-Audio-R1** ([HuggingFace](https://huggingface.co/stepfun-ai/Step-Audio-R1); [ModelScope](https://modelscope.cn/models/stepfun-ai/Step-Audio-R1))\n* Nov 27, 2025: 🎮 We released the [HF Space Playground](https://www.stepfun.com/studio/audio?tab=conversation)\n* Nov 19, 2025: 🎉 We release the [Demo Page](https://modelscope.cn/studios/stepfun-ai/Step-Audio-R1)\n* Nov 19, 2025: 👋 We rele"},{"ref":"P20","kind":"page","title":"stepfun-ai/stepfunApi-audio-sdk repository metadata","date":"2026-06-11T03:20:32.293529+00:00","date_source":null,"source_url":"https://github.com/stepfun-ai/stepfunApi-audio-sdk","signal_url":null,"signal_json_url":null,"text":"# stepfun-ai/stepfunApi-audio-sdk\n\nDescription: 基于阶跃星辰开放平台语音api的android 语音sdk，支持tts 流式与非流式，asr，流式，非流式音频播放器，语音录制能力\n\nLanguage: Kotlin\n\nLicense: Apache-2.0\n\nStars: 5\n\nForks: 1\n\nOpen issues: 0\n\nCreated: 2025-11-20T06:21:54Z\n\nPushed: 2025-12-15T08:27:36Z\n\nDefault branch: main\n\nFork: no\n\nArchived: no\n\nREADME:\n# StepFun Audio SDK 使用文档\n\n---\n\n## 目录\n\n1. [快速开始](#快速开始)\n2. [SDK 初始化](#sdk-初始化)\n3. [TTS 文字转语音](#tts-文字转语音)\n- [一次性生成](#一次性生成)\n- [生成并播放](#生成并播放)\n- [流式 TTS](#流式-tts)\n4. [ASR 语音识别](#asr-语音识别)\n- [识别音频文件](#识别音频文件)\n- [录音并识别](#录音并识别)\n5. [配置参数说明](#配置参数说明)\n6. [错误处理](#错误处理)\n7. [资源释放](#资源释放)\n\n---\n\n## 快速开始\n\n### 添加依赖\n\n```kotlin\nimplementation(\"com.stepfun:stepfun-audio-sdk:x.x.x\")\n```\n\n### 最简示例\n\n```kotlin\n// 1. 初始化 SDK\nSpeechSdk.init(context, SpeechConfig.Builder()\n.httpBaseUrl(\"https://your-api-server.com\")\n.webSocketUrl(\"wss://your-ws-server.com\")\n.customHeaders(mapOf(\"Authorization\" to \"Bearer YOUR_API_KEY\"))\n.enableLogging(true)\n.build()\n)\n\n// 2. 一键播放语音\nSpeechSdk.TTS.quickPlay(context, \"你好，世界！\")\n```\n\n---\n\n## SDK 初始化\n\n### SpeechConfig 配置\n\n```kotlin\nval config = SpeechConfig.Builder()\n// 服务器地址（必填）\n.httpBaseUrl(\"https://api.stepfun.com\")\n.webSocketUrl(\"wss://ws.stepfun.com\")\n\n// 认证信息（通过 headers 传递）\n.customHeaders(mapOf(\n\"Authorization\" to \"Bearer sk-xxxxxx\"\n))\n\n// 环境配置\n.environment(Environment.PRODUCTION)\n\n// 日志开关\n.enableLogging(true)\n\n// TTS 默认配置\n.ttsConfig(TtsConfig(\ndefaultModel = TtsModel.STEP_TTS_MINI.modelId,\ndefaultVoice = TtsVoice.STEP_TTS_MINI_DEFAULT,\ndefaultResponseFormat = TtsAudioFormat.PCM,\ndefaultSpeed = 1.0f,\ndefaultVolume = 1.0f,\ndefaultSampleRate = 24000\n))\n\n.build()\n\nSpeechSdk.init(context, config)\n```\n\n### 检查初始化状态\n\n```kotlin\nif (SpeechSdk.isInitialized()) {\n// SDK 已初始化\n}\n```\n\n### 更新配置\n\n```kotlin\nSpeechSdk.updateConfig(newConfig)\n```\n\n---\n\n## TTS 文字转语音\n\n### 一次性生成\n\n#### 使用默认配置\n\n```kotlin\nSpeechSdk.TTS.generateSpeech(\ntext = \"你好，我是阶跃星辰语音助手\",\nvoice = TtsVoice.STEP_TTS_MINI_DEFAULT,\ncallback = object : TtsCallback {\noverride fun onSuccess(audioData: ByteArray) {\n// 获取到音频数据（可保存或自行播放）\n}\n\noverride fun onError(error: TtsError) {\nLog.e(\"TTS\", \"生成失败: ${error.message}\")\n}\n}\n)\n```\n\n#### 使用自定义参数\n\n```kotlin\nval params = TtsSpeechParams.Builder()\n.model(TtsModel.STEP_TTS_MI"},{"ref":"P21","kind":"page","title":"stepfun-ai/Step-Audio-Edit-Benchmark repository metadata","date":"2026-06-11T03:20:32.275006+00:00","date_source":null,"source_url":"https://github.com/stepfun-ai/Step-Audio-Edit-Benchmark","signal_url":null,"signal_json_url":null,"text":"# stepfun-ai/Step-Audio-Edit-Benchmark\n\nLanguage: Python\n\nStars: 20\n\nForks: 1\n\nOpen issues: 0\n\nCreated: 2025-11-20T06:36:16Z\n\nPushed: 2025-11-22T15:18:25Z\n\nDefault branch: main\n\nFork: no\n\nArchived: no\n\nREADME:\n# Step-Audio-Edit-Benchmark\n\n<p align=\"center\">\n<img src=\"assets/logo.png\" height=100>\n</p>\n\n# Introduction\n\nEvaluating controllable speech synthesis remains challenging due to the lack of a single comprehensive benchmark that simultaneously covers fine-grained attributes like emotion, speaking style, and paralinguistics.\n\nWe introduce **step-audio-edit-benchmark**, a comprehensive evaluation framework covering emotion, speaking style, and paralinguistics, as introduced in technical report [Step-Audio-EditX](https://arxiv.org/abs/2511.03601).\n\n# Dataset\n## Prompt Audios\n\nWe selected 8 speakers in total (4 Chinese and 4 English), balanced with two males and two females per language. The Chinese data is sourced from [WenetSpeech4TTS](https://huggingface.co/datasets/Wenetspeech4TTS/WenetSpeech4TTS), while the English data comes from [GLOBE_V2](https://huggingface.co/datasets/MushanW/GLOBE_V2) and [Libri-Light](https://github.com/facebookresearch/libri-light).\n\nAdditionally, we provide the long audio samples referenced in our paper [Step-Audio-EditX](https://arxiv.org/abs/2511.03601), designed for evaluating the voice cloning capabilities of closed-source models.\n\n<div align=\"center\">\n\n<table>\n<thead>\n<tr>\n<th align=\"center\">lang</th>\n<th align=\"center\">speaker</th>\n<th align=\"center\">gender</th>\n<th align=\"center\">from</th>\n</tr>\n</thead>\n<tbody>\n<tr>\n<td rowspan=\"4\" align=\"center\">zh</td>\n<td align=\"center\">Y0000004339_A_SMNK0c4uM_S00403-S00406</td>\n<td align=\"center\">Female</td>\n<td align=\"center\">WenetSpeech4TTS</td>\n</tr>\n<tr>\n<td align=\"center\">X0000015410_331546220_S00073-S00074</td>\n<td align=\"center\">Female</td>\n<td align=\"center\">WenetSpeech4TTS</td>\n</tr>\n<tr>\n<td align=\"center\">X0000005119_6330761_S01227-S01229</td>\n<td align=\"center\">Male</td>\n<td align=\"center\">WenetSpeech4TTS</td>\n</tr>\n<tr>\n<td align=\"center\">X0000000863_279853194_S00611-S00612</td>\n<td align=\"center\">Male</td>\n<td align=\"center\">WenetSpeech4TTS</td>\n</tr>\n<tr>\n<td rowspan=\"4\" "},{"ref":"P22","kind":"page","title":"stepfun-ai/gelab-zero repository metadata","date":"2026-06-11T03:20:31.882289+00:00","date_source":null,"source_url":"https://github.com/stepfun-ai/gelab-zero","signal_url":null,"signal_json_url":null,"text":"# stepfun-ai/gelab-zero\n\nDescription: STEP-GUI: The top GUI agent solution in the galaxy. Developed by the StepFun-GELab team and powered by StepFun’s cutting-edge research capabilities.\n\nLanguage: Python\n\nLicense: MIT\n\nStars: 2191\n\nForks: 193\n\nOpen issues: 44\n\nCreated: 2025-11-28T14:42:44Z\n\nPushed: 2026-05-11T05:50:07Z\n\nDefault branch: main\n\nFork: no\n\nArchived: no\n\nREADME:\n![GELab-Zero Main Image](./images/main_en.png)\n\n> 👋 Hi, everyone! We are proud to present the first fully open-source GUI Agent with both model and infrastructure. Our solution features plug-and-play engineering with no cloud dependencies, giving you complete privacy control.\n\n<p align=\"center\">\n<!-- <a href=\"https://github.com/stepfun-ai/gelab-zero\"><img src=\"https://img.shields.io/badge/💻%20GitHub-Repository-black\" alt=\"GitHub\" /></a> -->\n<a href=\"https://opengelab.github.io/\"><img src=\"https://img.shields.io/badge/🌐%20Website-Project%20Page-blue\" alt=\"Website\" /></a>\n<a href=\"https://huggingface.co/stepfun-ai/GELab-Zero-4B-preview\"><img src=\"https://img.shields.io/badge/🤗%20Hugging%20Face-GELab--Zero--4B--preview-orange\" alt=\"Hugging Face Model\" /></a>\n<a href=\"https://huggingface.co/datasets/stepfun-ai/AndroidDaily\"><img src=\"https://img.shields.io/badge/📚%20Hugging%20Face-AndroidDaily-yellow\" alt=\"Hugging Face Dataset\" /></a>\n<a href=\"https://modelscope.cn/models/stepfun-ai/GELab-Zero-4B-preview\"><img src=\"https://img.shields.io/badge/🤖%20Model%20Scope-GELab--Zero--4B--preview-blue\" alt=\"Model Scope\" /></a>\n</p>\n\n<p align=\"center\">\n<a href=\"./README.md\">English</a> |\n<a href=\"./README_CN.md\">简体中文</a>\n</p>\n\n## 📰 News\n\n* 🎁 **[Coming Soon...]**\n\n* 🎁 **[2025-12-12]** MCP-Server ready：\n\n<!-- ### Step1 启动 mcp server 以支持多设备管理和任务分发 -->\n### Step1 Start MCP server to support multi-device management and task distribution\n\n```bash\n# enable mcp server\npython mcp_server/detailed_gelab_mcp_server.py\n```\n\n### Step2 Import MCP tools in Chatbox\n<!-- images/MCP-chatbox.png -->\n<div style=\"display: flex; align-items: center; justify-content: center; width: 80%; margin: 0 auto;\">\n<img src=\"images/MCP-chatbox.png\" alt=\"MCP-Demo\" style=\"flex: 1; height: 400px; object-fit: contain; margin-right: 1px;\"/"},{"ref":"P23","kind":"page","title":"stepfun-ai/StepDeepResearch repository metadata","date":"2026-06-11T03:20:31.822221+00:00","date_source":null,"source_url":"https://github.com/stepfun-ai/StepDeepResearch","signal_url":null,"signal_json_url":null,"text":"# stepfun-ai/StepDeepResearch\n\nDescription: Step-DeepResearch\n\nLanguage: Python\n\nLicense: Apache-2.0\n\nStars: 561\n\nForks: 26\n\nOpen issues: 3\n\nCreated: 2025-11-26T06:02:31Z\n\nPushed: 2026-03-24T07:08:07Z\n\nDefault branch: main\n\nFork: no\n\nArchived: no\n\nREADME:\n# Step-DeepResearch\n\n<div align=\"center\">\n<img src=\"assets/logo.png\" height=100>\n</div>\n<div align=\"center\" style=\"line-height: 1;\">\n<a href=\"https://stepfun.com/\" target=\"_blank\"><img alt=\"Homepage\" src=\"https://img.shields.io/badge/Homepage-StepFun-white?logo=StepFun&logoColor=white\"/></a> &ensp;\n<a href=\"https://x.com/StepFun_ai\" target=\"_blank\"><img alt=\"Twitter Follow\" src=\"https://img.shields.io/badge/Twitter-StepFun-white?logo=x&logoColor=white\"/></a> &ensp;\n<a href=\"https://discord.com/invite/XHheP5Fn\" target=\"_blank\"><img alt=\"Discord\" src=\"https://img.shields.io/badge/Discord-StepFun-white?logo=discord&logoColor=white\"/></a>\n</div>\n<div align=\"center\">\n<a href=\"https://arxiv.org/pdf/2512.20491\"><img src=\"https://img.shields.io/static/v1?label=Step-DeepResearch&message=Arxiv&color=red\"></a> &ensp;\n<a href=\"https://platform.stepfun.com/interface-key\"><img src=\"https://img.shields.io/static/v1?label=Step-DeepResearch&message=Model%20API&color=blue\"></a>\n<a href=\"https://github.com/stepfun-ai/StepDeepResearch/blob/main/LICENSE\"><img alt=\"License\" src=\"https://img.shields.io/badge/License-Apache%202.0-blue?&color=blue\"/></a>\n</div>\n\n## News\n\n* Feb 2, 2026: 👋 We have released **Step 3.5 Flash**, achieving **65.27** on <span style=\"font-variant: small-caps;\">ResearchRubrics</span>. Try it out by setting the environment variable `MODEL_NAME=step-3.5-flash`. [Details](https://static.stepfun.com/blog/step-3.5-flash/)\n\n* Dec 25, 2025: 👋 You can join our group chat to get updates on your beta API application status and the latest project developments.\n<div align=\"center\">\n<img src=\"assets/wechat_qr_code.jpg\" alt=\"WeChat QR code\" width=\"180\" />\n<img src=\"assets/feishu_qr_code.png\" alt=\"Feishu QR code\" width=\"180\" />\n</div>\n\n* Dec 24, 2025: 👋 We have made our technical report available. [Read](https://arxiv.org/pdf/2512.20491)\n\n## Introduction\n### Model Summary\n**Step-DeepResearch** is a cost-effective, end-to-e"},{"ref":"P24","kind":"page","title":"stepfun-ai/PaCoRe repository metadata","date":"2026-06-11T03:20:31.707947+00:00","date_source":null,"source_url":"https://github.com/stepfun-ai/PaCoRe","signal_url":null,"signal_json_url":null,"text":"# stepfun-ai/PaCoRe\n\nDescription: PaCoRe: Learning to Scale Test-Time Compute with Parallel Coordinated Reasoning\n\nLanguage: Python\n\nStars: 334\n\nForks: 15\n\nOpen issues: 0\n\nCreated: 2025-12-09T10:43:09Z\n\nPushed: 2026-02-05T01:27:48Z\n\nDefault branch: main\n\nFork: no\n\nArchived: no\n\nREADME:\n# PaCoRe: Learning to Scale Test-Time Compute with Parallel Coordinated Reasoning\n\n<div align=\"center\">\n\n[**Read the Paper**](https://arxiv.org/abs/2601.05593) | [**Download Models**](https://huggingface.co/stepfun-ai/PaCoRe-8B) | [**Training Data**](https://huggingface.co/datasets/stepfun-ai/PaCoRe-Train-8k)\n\n</div>\n\n## 📖 Overview\n\nWe introduce **PaCoRe (Parallel Coordinated Reasoning)**, a framework that shifts the driver of inference from sequential depth to **coordinated parallel breadth**, breaking the model context limitation and massively scaling test time compute: \n* **Think in Parallel:** PaCoRe launches massive parallel exploration trajectories.\n* **Coordinate in Multi-rounds:** It employs a message-passing architecture to compact these thoughts into concise messages and synthesize them to guide the next round.\n\nTrained via large-scale, outcome-based reinforcement learning, PaCoRe masters the **Reasoning Synthesis** capabilities required to reconcile diverse parallel insights. \n\nThe approach yields strong improvements across diverse domains, and notably pushes reasoning beyond frontier systems in mathematics: an 8B model reaches 94.5\\% on HMMT 2025, surpassing GPT-5’s 93.2\\% by scaling effective TTC to roughly two million tokens.\n\nWe open-source model checkpoints, training data, and the full inference pipeline to accelerate follow-up work!\n\n------\n\n<p align=\"center\">\n<img src=\"figure/teaser_draft_02.png\" width=\"48%\" />\n<img src=\"figure/before_after_train_lcb_02.png\" width=\"48%\" />\n</p>\n\n*Figure 1 | Parallel Coordinated Reasoning (PaCoRe) performance. Left: On HMMT 2025, PaCoRe-8B demonstrates remarkable test-time scaling, yielding steady gains and ultimately surpassing GPT-5. Right: On LiveCodeBench, the RLVR-8B model fails to leverage increased test-time compute, while PaCoRe-8B model effectively unlocks substantial gains as the test-time compute increases.*\n\n<p align="},{"ref":"P25","kind":"page","title":"stepfun-ai/Step1X-3D v.1.0.0","date":"2026-06-11T03:17:12.07265+00:00","date_source":null,"source_url":"https://github.com/stepfun-ai/Step1X-3D/releases/tag/v.1.0.0","signal_url":null,"signal_json_url":null,"text":"# Step1X-3D-v.1.0.0\n\nRepository: stepfun-ai/Step1X-3D\n\nTag: v.1.0.0\n\nPublished: 2025-05-13T03:53:20Z\n\nPrerelease: no\n\nRelease notes: none published."},{"ref":"P26","kind":"page","title":"stepfun-ai/Step1X-3D v.1.0.1","date":"2026-06-11T03:17:11.958135+00:00","date_source":null,"source_url":"https://github.com/stepfun-ai/Step1X-3D/releases/tag/v.1.0.1","signal_url":null,"signal_json_url":null,"text":"# Step1X-3D v.1.0.1\n\nRepository: stepfun-ai/Step1X-3D\n\nTag: v.1.0.1\n\nPublished: 2025-05-14T08:25:33Z\n\nPrerelease: no\n\nRelease notes:\n1. Fix gradio app load JPEG error\n2. Fix symmetry label process error\n3. Improve readme.md"},{"ref":"P27","kind":"page","title":"stepfun-ai/SteptronOss repository metadata","date":"2026-06-11T03:03:15.060901+00:00","date_source":null,"source_url":"https://github.com/stepfun-ai/SteptronOss","signal_url":null,"signal_json_url":null,"text":"# stepfun-ai/SteptronOss\n\nDescription: A lightweight, AI-native training framework for large language models. Designed for fast iteration, reproducible experiments, and modular configuration across SFT, RLVR, and evaluation workflows.\n\nLanguage: Python\n\nLicense: Apache-2.0\n\nStars: 575\n\nForks: 43\n\nOpen issues: 8\n\nCreated: 2025-12-18T11:58:05Z\n\nPushed: 2026-05-18T07:15:55Z\n\nDefault branch: dev\n\nFork: no\n\nArchived: no\n\nREADME:\n<div align=\"center\">\n<h1 style=\"margin: 0; border-bottom: none;\"> <img src=\"assets/stepfun.svg\" width=\"25\" style=\"margin-right: 10px;\"/> StepTronOSS </h1>\n</div>\n\n<p align=\"center\">\n<strong>English</strong>&nbsp; | &nbsp;<a href=\"README_ZH.md\">简体中文</a>\n</p>\n\n**StepTronOSS** is a lightweight **training framework** for large-scale language models with a focus on modular configs, reproducible experiments, and fast iteration across SFT, RLVR, and evaluation workflows. It can run with only PyTorch as a dependency, while also supporting operator-level replacements for acceleration.\n\n## Key capabilities:\n- Config-driven experiments with dynamic validation (`cfshow`, `sanity_check`)\n- Multi-task orchestration with flexible launch tooling\n- Extensible data/optimizer/model stacks for rapid research iteration\n\n## Docs\n\n- Launch guide (EN): `docs/LAUNCH_EXPERIMENTS.md`\n- Launch guide (ZH): `docs/LAUNCH_EXPERIMENTS_ZH.md`\n- SFT data prep (ZH): `docs/SFT_DATA_PREPARATION.md`\n- SFT data prep (EN): `docs/SFT_DATA_PREPARATION_EN.md`\n- API modules: `docs/MODULES.md`\n\n## Installation\n\n```bash\n# from repo root\nuv sync\n# install redis-server\napt install -y redis-server\n```\n\n## Getting Started\n\nuv virtual environment is recommanded. If not, prefix with `uv run`.\n\n### Experiment Overview & Sanity Check\n\n```bash\n# Overview the experiment config and run sanity_check\nuv run cfshow playground/rlvr/qwen3_1p5b_rlvr_math.py\n# Inspect a specific subtree (e.g., actor config)\nuv run cfshow playground/rlvr/qwen3_1p5b_rlvr_math.py -k actor_model_cfg\n```\n\n### Run Experiments\n\n```bash\n# Single-task experiments (e.g., SFT)\nuv run torchrun playground/sft/your_exp.py\n\n# Multi-task experiments (e.g., RL)\nexport STEPTRON_MEET_DIR=/path/to/shared\nuv run tools/mp_run.py playground/rlvr"},{"ref":"P28","kind":"page","title":"stepfun-ai/Step3-VL-10B repository metadata","date":"2026-06-11T03:02:54.090598+00:00","date_source":null,"source_url":"https://github.com/stepfun-ai/Step3-VL-10B","signal_url":null,"signal_json_url":null,"text":"# stepfun-ai/Step3-VL-10B\n\nDescription: Step3-VL-10B: A compact yet frontier multimodal model achieving SOTA performance at the 10B scale, matching open-source models 10-20x its size.\n\nLicense: Apache-2.0\n\nStars: 407\n\nForks: 30\n\nOpen issues: 19\n\nCreated: 2026-01-13T09:13:12Z\n\nPushed: 2026-01-21T13:42:15Z\n\nDefault branch: main\n\nFork: no\n\nArchived: no\n\nREADME:\n<div align=\"center\">\n\n<div align=\"center\">\n<h1 style=\"border-bottom: none;\">\n<img src=\"figures/stepfun.svg\" width=\"30\" style=\"vertical-align: bottom; margin-right: 10px;\" /> \nSTEP3-VL-10B\n</h1>\n</div>\n\n[![Hugging Face](https://img.shields.io/badge/%F0%9F%A4%97%20HF-StepFun/STEP3--VL--10B-blue)](https://huggingface.co/collections/stepfun-ai/step3-vl-10b)\n[![ModelScope](https://img.shields.io/badge/ModelScope-StepFun/STEP3--VL--10B-624aff)](https://modelscope.cn/collections/stepfun-ai/Step3-VL-10B)\n[![Paper](https://img.shields.io/badge/Paper-Arxiv-red)](https://arxiv.org/abs/2601.09668)\n[![License](https://img.shields.io/badge/License-Apache%202.0-green)]()\n\n</div>\n\n## 🚀 Introduction\n\n**STEP3-VL-10B** is a lightweight open-source foundation model designed to redefine the trade-off between compact efficiency and frontier-level multimodal intelligence. Despite its compact **10B parameter footprint**, STEP3-VL-10B excels in **visual perception**, **complex reasoning**, and **human-centric alignment**. It consistently outperforms models under the 10B scale and rivals or surpasses significantly larger open-weights models (**10×–20× its size**), such as GLM-4.6V (106B-A12B), Qwen3-VL-Thinking (235B-A22B), and top-tier proprietary flagships like Gemini 2.5 Pro and Seed-1.5-VL.\n\n<div align=\"center\">\n<img src=\"figures/performance.png\" alt=\"Performance Comparison\" width=\"800\"/>\n<p><i>Figure 1: Performance comparison of STEP3-VL-10B against SOTA multimodal foundation models. SeRe: Sequential Reasoning; PaCoRe: Parallel Coordinated Reasoning.</i></p>\n</div>\n\nThe success of STEP3-VL-10B is driven by two key strategic designs:\n\n1. **Unified Pre-training on High-Quality Multimodal Corpus:** A single-stage, fully unfrozen training strategy on a 1.2T token multimodal corpus, focusing on two foundational capabilities: **reaso"},{"ref":"E1","kind":"event","title":"stepfun-ai/Step-3.5-Flash","date":"2026-02-01T08:03:45+00:00","date_source":"source","source_url":"https://huggingface.co/stepfun-ai/Step-3.5-Flash","signal_url":"https://onlylabs.fyi/signals/8dfff472-9670-4ad2-b0b3-3e52cad3178f","signal_json_url":"https://onlylabs.fyi/signals/8dfff472-9670-4ad2-b0b3-3e52cad3178f/signal.json","text":"model_released · stepfun-ai/Step-3.5-Flash · signal_desk=releases · occurred_at=2026-02-01T08:03:45+00:00 · url=https://huggingface.co/stepfun-ai/Step-3.5-Flash · hf_downloads=325856 · hf_likes=820 · hf_params=199384301376 · pipeline=text-generation · license=apache-2.0"},{"ref":"E2","kind":"event","title":"stepfun-ai/Step3-VL-10B","date":"2026-01-13T07:55:22+00:00","date_source":"source","source_url":"https://huggingface.co/stepfun-ai/Step3-VL-10B","signal_url":"https://onlylabs.fyi/signals/51dcd072-6a22-4f0e-aa59-7fa5ab0decf2","signal_json_url":"https://onlylabs.fyi/signals/51dcd072-6a22-4f0e-aa59-7fa5ab0decf2/signal.json","text":"model_released · stepfun-ai/Step3-VL-10B · signal_desk=releases · occurred_at=2026-01-13T07:55:22+00:00 · url=https://huggingface.co/stepfun-ai/Step3-VL-10B · hf_downloads=484259 · hf_likes=406 · hf_params=10171750144 · pipeline=image-text-to-text · license=apache-2.0 · raw={\"derived_reason\":\"first-party-finetune\"}"},{"ref":"E3","kind":"event","title":"stepfun-ai/Step-3.7-Flash","date":"2026-05-23T02:13:46+00:00","date_source":"source","source_url":"https://huggingface.co/stepfun-ai/Step-3.7-Flash","signal_url":"https://onlylabs.fyi/signals/75371011-2872-44b9-bf60-c41fb2e2f0fb","signal_json_url":"https://onlylabs.fyi/signals/75371011-2872-44b9-bf60-c41fb2e2f0fb/signal.json","text":"model_released · stepfun-ai/Step-3.7-Flash · signal_desk=releases · occurred_at=2026-05-23T02:13:46+00:00 · url=https://huggingface.co/stepfun-ai/Step-3.7-Flash · hf_downloads=50187 · hf_likes=363 · hf_params=201365316160 · pipeline=image-text-to-text · license=apache-2.0"},{"ref":"E4","kind":"event","title":"stepfun-ai/Step-Audio-R1.1","date":"2026-01-14T06:37:24+00:00","date_source":"source","source_url":"https://huggingface.co/stepfun-ai/Step-Audio-R1.1","signal_url":"https://onlylabs.fyi/signals/32432450-8ccf-4dda-ac3b-cb7fee5caee8","signal_json_url":"https://onlylabs.fyi/signals/32432450-8ccf-4dda-ac3b-cb7fee5caee8/signal.json","text":"model_released · stepfun-ai/Step-Audio-R1.1 · signal_desk=releases · occurred_at=2026-01-14T06:37:24+00:00 · url=https://huggingface.co/stepfun-ai/Step-Audio-R1.1 · hf_downloads=178 · hf_likes=181 · hf_params=33487033600 · pipeline=audio-text-to-text · license=apache-2.0"},{"ref":"E5","kind":"event","title":"stepfun-ai/step3","date":"2025-07-28T11:07:33+00:00","date_source":"source","source_url":"https://huggingface.co/stepfun-ai/step3","signal_url":"https://onlylabs.fyi/signals/2bee57ac-0972-404c-933a-a5839630207a","signal_json_url":"https://onlylabs.fyi/signals/2bee57ac-0972-404c-933a-a5839630207a/signal.json","text":"model_released · stepfun-ai/step3 · signal_desk=releases · occurred_at=2025-07-28T11:07:33+00:00 · url=https://huggingface.co/stepfun-ai/step3 · hf_downloads=144095 · hf_likes=166 · hf_params=320969912832 · pipeline=image-text-to-text · license=apache-2.0"},{"ref":"E6","kind":"event","title":"stepfun-ai/GELab-Zero-4B-preview","date":"2025-11-28T15:00:26+00:00","date_source":"source","source_url":"https://huggingface.co/stepfun-ai/GELab-Zero-4B-preview","signal_url":"https://onlylabs.fyi/signals/50020ff7-8686-4212-b088-7e1f0bc425bc","signal_json_url":"https://onlylabs.fyi/signals/50020ff7-8686-4212-b088-7e1f0bc425bc/signal.json","text":"model_released · stepfun-ai/GELab-Zero-4B-preview · signal_desk=releases · occurred_at=2025-11-28T15:00:26+00:00 · url=https://huggingface.co/stepfun-ai/GELab-Zero-4B-preview · hf_downloads=2214 · hf_likes=152 · hf_params=4437815808 · pipeline=image-text-to-text · license=apache-2.0 · raw={\"derived_reason\":\"first-party-finetune\"}"},{"ref":"E7","kind":"event","title":"stepfun-ai/Step-Audio-R1","date":"2025-11-18T13:49:22+00:00","date_source":"source","source_url":"https://huggingface.co/stepfun-ai/Step-Audio-R1","signal_url":"https://onlylabs.fyi/signals/eb8e8d87-88b5-4a45-8c4c-66abc5cc1fba","signal_json_url":"https://onlylabs.fyi/signals/eb8e8d87-88b5-4a45-8c4c-66abc5cc1fba/signal.json","text":"model_released · stepfun-ai/Step-Audio-R1 · signal_desk=releases · occurred_at=2025-11-18T13:49:22+00:00 · url=https://huggingface.co/stepfun-ai/Step-Audio-R1 · hf_downloads=222 · hf_likes=144 · hf_params=33487033600 · pipeline=audio-text-to-text · license=apache-2.0"},{"ref":"E8","kind":"event","title":"stepfun-ai/Step1X-3D","date":"2025-05-11T15:20:00+00:00","date_source":"source","source_url":"https://huggingface.co/stepfun-ai/Step1X-3D","signal_url":"https://onlylabs.fyi/signals/effb934f-2d25-4852-a215-81295098dc18","signal_json_url":"https://onlylabs.fyi/signals/effb934f-2d25-4852-a215-81295098dc18/signal.json","text":"model_released · stepfun-ai/Step1X-3D · signal_desk=releases · occurred_at=2025-05-11T15:20:00+00:00 · url=https://huggingface.co/stepfun-ai/Step1X-3D · hf_likes=107 · license=apache-2.0"},{"ref":"E9","kind":"event","title":"stepfun-ai/NextStep-1-Large","date":"2025-08-12T16:52:03+00:00","date_source":"source","source_url":"https://huggingface.co/stepfun-ai/NextStep-1-Large","signal_url":"https://onlylabs.fyi/signals/661477dc-daa6-4fcc-a715-64ba0019fb6f","signal_json_url":"https://onlylabs.fyi/signals/661477dc-daa6-4fcc-a715-64ba0019fb6f/signal.json","text":"model_released · stepfun-ai/NextStep-1-Large · signal_desk=releases · occurred_at=2025-08-12T16:52:03+00:00 · url=https://huggingface.co/stepfun-ai/NextStep-1-Large · hf_downloads=62 · hf_likes=98 · hf_params=14959056960 · pipeline=text-to-image · license=apache-2.0"},{"ref":"E10","kind":"event","title":"stepfun-ai/Step-3.5-Flash-Base","date":"2026-03-02T16:17:50+00:00","date_source":"source","source_url":"https://huggingface.co/stepfun-ai/Step-3.5-Flash-Base","signal_url":"https://onlylabs.fyi/signals/1a84a279-2a4f-4253-80c9-4f07acd82b3f","signal_json_url":"https://onlylabs.fyi/signals/1a84a279-2a4f-4253-80c9-4f07acd82b3f/signal.json","text":"model_released · stepfun-ai/Step-3.5-Flash-Base · signal_desk=releases · occurred_at=2026-03-02T16:17:50+00:00 · url=https://huggingface.co/stepfun-ai/Step-3.5-Flash-Base · hf_downloads=140 · hf_likes=84 · hf_params=197764857152 · pipeline=text-generation · license=apache-2.0"},{"ref":"E11","kind":"event","title":"stepfun-ai/Step1X-Edit-v1p2","date":"2025-11-26T06:12:18+00:00","date_source":"source","source_url":"https://huggingface.co/stepfun-ai/Step1X-Edit-v1p2","signal_url":"https://onlylabs.fyi/signals/0842b075-b280-4499-a41e-710c72687b83","signal_json_url":"https://onlylabs.fyi/signals/0842b075-b280-4499-a41e-710c72687b83/signal.json","text":"model_released · stepfun-ai/Step1X-Edit-v1p2 · signal_desk=releases · occurred_at=2025-11-26T06:12:18+00:00 · url=https://huggingface.co/stepfun-ai/Step1X-Edit-v1p2 · hf_downloads=530 · hf_likes=61 · pipeline=image-to-image · license=apache-2.0"},{"ref":"E12","kind":"event","title":"stepfun-ai/Step3-VL-10B-Base","date":"2026-01-13T12:04:54+00:00","date_source":"source","source_url":"https://huggingface.co/stepfun-ai/Step3-VL-10B-Base","signal_url":"https://onlylabs.fyi/signals/14e3776b-ee83-412b-ae9c-5f3ab0007021","signal_json_url":"https://onlylabs.fyi/signals/14e3776b-ee83-412b-ae9c-5f3ab0007021/signal.json","text":"model_released · stepfun-ai/Step3-VL-10B-Base · signal_desk=releases · occurred_at=2026-01-13T12:04:54+00:00 · url=https://huggingface.co/stepfun-ai/Step3-VL-10B-Base · hf_downloads=97 · hf_likes=51 · hf_params=10171750144 · pipeline=image-text-to-text · license=apache-2.0"},{"ref":"E13","kind":"event","title":"stepfun-ai/NextStep-1-Large-Edit","date":"2025-08-12T02:57:31+00:00","date_source":"source","source_url":"https://huggingface.co/stepfun-ai/NextStep-1-Large-Edit","signal_url":"https://onlylabs.fyi/signals/a3341e22-6fcd-43d9-9c2b-f7bc804e44cb","signal_json_url":"https://onlylabs.fyi/signals/a3341e22-6fcd-43d9-9c2b-f7bc804e44cb/signal.json","text":"model_released · stepfun-ai/NextStep-1-Large-Edit · signal_desk=releases · occurred_at=2025-08-12T02:57:31+00:00 · url=https://huggingface.co/stepfun-ai/NextStep-1-Large-Edit · hf_downloads=44 · hf_likes=50 · hf_params=14953814080 · pipeline=image-to-image · license=apache-2.0"},{"ref":"E14","kind":"event","title":"stepfun-ai/Step-Audio-AQAA","date":"2025-06-11T12:09:31+00:00","date_source":"source","source_url":"https://huggingface.co/stepfun-ai/Step-Audio-AQAA","signal_url":"https://onlylabs.fyi/signals/0dddf78b-0dbc-4f79-82c3-859d99f853d1","signal_json_url":"https://onlylabs.fyi/signals/0dddf78b-0dbc-4f79-82c3-859d99f853d1/signal.json","text":"model_released · stepfun-ai/Step-Audio-AQAA · signal_desk=releases · occurred_at=2025-06-11T12:09:31+00:00 · url=https://huggingface.co/stepfun-ai/Step-Audio-AQAA · hf_downloads=38 · hf_likes=49 · hf_params=136663907840 · license=apache-2.0"},{"ref":"E15","kind":"event","title":"stepfun-ai/PaCoRe-8B","date":"2025-12-09T06:52:38+00:00","date_source":"source","source_url":"https://huggingface.co/stepfun-ai/PaCoRe-8B","signal_url":"https://onlylabs.fyi/signals/da4b930b-d894-4edc-b181-d273e14a9a13","signal_json_url":"https://onlylabs.fyi/signals/da4b930b-d894-4edc-b181-d273e14a9a13/signal.json","text":"model_released · stepfun-ai/PaCoRe-8B · signal_desk=releases · occurred_at=2025-12-09T06:52:38+00:00 · url=https://huggingface.co/stepfun-ai/PaCoRe-8B · hf_downloads=110 · hf_likes=48 · pipeline=text-generation · license=mit"},{"ref":"E16","kind":"event","title":"stepfun-ai/Step-Realtime-CLI v0.1.0","date":"2026-06-11T06:32:05+00:00","date_source":"source","source_url":"https://github.com/stepfun-ai/Step-Realtime-CLI/releases/tag/v0.1.0","signal_url":"https://onlylabs.fyi/signals/5d96e475-3f22-48f3-825b-524616206359","signal_json_url":"https://onlylabs.fyi/signals/5d96e475-3f22-48f3-825b-524616206359/signal.json","text":"release · stepfun-ai/Step-Realtime-CLI v0.1.0 · signal_desk=releases · occurred_at=2026-06-11T06:32:05+00:00 · url=https://github.com/stepfun-ai/Step-Realtime-CLI/releases/tag/v0.1.0 · raw={\"repo\":\"stepfun-ai/Step-Realtime-CLI\"}"},{"ref":"E17","kind":"event","title":"stepfun-ai/Step-3.5-Flash-Base-Midtrain","date":"2026-03-02T09:49:21+00:00","date_source":"source","source_url":"https://huggingface.co/stepfun-ai/Step-3.5-Flash-Base-Midtrain","signal_url":"https://onlylabs.fyi/signals/65659b12-3d4e-4cb7-96f5-783dc1bc2372","signal_json_url":"https://onlylabs.fyi/signals/65659b12-3d4e-4cb7-96f5-783dc1bc2372/signal.json","text":"model_released · stepfun-ai/Step-3.5-Flash-Base-Midtrain · signal_desk=releases · occurred_at=2026-03-02T09:49:21+00:00 · url=https://huggingface.co/stepfun-ai/Step-3.5-Flash-Base-Midtrain · hf_downloads=88 · hf_likes=42 · hf_params=197765520704 · pipeline=text-generation · license=apache-2.0"},{"ref":"E18","kind":"event","title":"stepfun-ai/Step-Realtime-CLI","date":"2026-06-01T06:00:45+00:00","date_source":"source","source_url":"https://github.com/stepfun-ai/Step-Realtime-CLI","signal_url":"https://onlylabs.fyi/signals/90ab2903-1de5-475c-a9f0-3fc16ca5bbf4","signal_json_url":"https://onlylabs.fyi/signals/90ab2903-1de5-475c-a9f0-3fc16ca5bbf4/signal.json","text":"repo_new · stepfun-ai/Step-Realtime-CLI · signal_desk=repos · occurred_at=2026-06-01T06:00:45+00:00 · url=https://github.com/stepfun-ai/Step-Realtime-CLI · stars=18 · raw={\"repo\":\"stepfun-ai/Step-Realtime-CLI\",\"language\":\"TypeScript\"}"},{"ref":"E19","kind":"event","title":"stepfun-ai/Step-3.7-Flash","date":"2026-05-27T09:06:25+00:00","date_source":"source","source_url":"https://github.com/stepfun-ai/Step-3.7-Flash","signal_url":"https://onlylabs.fyi/signals/a6e66305-4aef-4f71-88bb-ae062e216220","signal_json_url":"https://onlylabs.fyi/signals/a6e66305-4aef-4f71-88bb-ae062e216220/signal.json","text":"repo_new · stepfun-ai/Step-3.7-Flash · signal_desk=repos · occurred_at=2026-05-27T09:06:25+00:00 · url=https://github.com/stepfun-ai/Step-3.7-Flash · stars=228 · raw={\"repo\":\"stepfun-ai/Step-3.7-Flash\"}"},{"ref":"E20","kind":"event","title":"stepfun-ai/NextStep-1.1","date":"2025-12-23T15:26:53+00:00","date_source":"source","source_url":"https://huggingface.co/stepfun-ai/NextStep-1.1","signal_url":"https://onlylabs.fyi/signals/02e5c6f7-c2c1-46a1-b92d-5e9470c3db24","signal_json_url":"https://onlylabs.fyi/signals/02e5c6f7-c2c1-46a1-b92d-5e9470c3db24/signal.json","text":"model_released · stepfun-ai/NextStep-1.1 · signal_desk=releases · occurred_at=2025-12-23T15:26:53+00:00 · url=https://huggingface.co/stepfun-ai/NextStep-1.1 · hf_downloads=6596 · hf_likes=28 · hf_params=14953814080 · pipeline=text-to-image · license=apache-2.0"},{"ref":"E21","kind":"event","title":"stepfun-ai/NextStep-1-Large-Pretrain","date":"2025-08-15T08:34:42+00:00","date_source":"source","source_url":"https://huggingface.co/stepfun-ai/NextStep-1-Large-Pretrain","signal_url":"https://onlylabs.fyi/signals/861f4ba1-4c30-423c-8068-67cc9006918f","signal_json_url":"https://onlylabs.fyi/signals/861f4ba1-4c30-423c-8068-67cc9006918f/signal.json","text":"model_released · stepfun-ai/NextStep-1-Large-Pretrain · signal_desk=releases · occurred_at=2025-08-15T08:34:42+00:00 · url=https://huggingface.co/stepfun-ai/NextStep-1-Large-Pretrain · hf_downloads=9 · hf_likes=18 · hf_params=14953814080 · pipeline=text-to-image · license=apache-2.0"},{"ref":"E22","kind":"event","title":"stepfun-ai/Step1X-Edit-v1p2-preview","date":"2025-09-07T12:45:47+00:00","date_source":"source","source_url":"https://huggingface.co/stepfun-ai/Step1X-Edit-v1p2-preview","signal_url":"https://onlylabs.fyi/signals/f5803a7b-1afd-466f-a663-c666c8361806","signal_json_url":"https://onlylabs.fyi/signals/f5803a7b-1afd-466f-a663-c666c8361806/signal.json","text":"model_released · stepfun-ai/Step1X-Edit-v1p2-preview · signal_desk=releases · occurred_at=2025-09-07T12:45:47+00:00 · url=https://huggingface.co/stepfun-ai/Step1X-Edit-v1p2-preview · hf_downloads=11 · hf_likes=17 · pipeline=image-to-image · license=apache-2.0"},{"ref":"E23","kind":"event","title":"stepfun-ai/Step-Video-T2V","date":"2025-02-08T08:46:51+00:00","date_source":"source","source_url":"https://github.com/stepfun-ai/Step-Video-T2V","signal_url":"https://onlylabs.fyi/signals/de87612c-23b1-4a49-999f-f3c7f6a3accf","signal_json_url":"https://onlylabs.fyi/signals/de87612c-23b1-4a49-999f-f3c7f6a3accf/signal.json","text":"repo_new · stepfun-ai/Step-Video-T2V · signal_desk=repos · occurred_at=2025-02-08T08:46:51+00:00 · url=https://github.com/stepfun-ai/Step-Video-T2V · stars=3185 · raw={\"repo\":\"stepfun-ai/Step-Video-T2V\",\"language\":\"Python\"}"},{"ref":"E24","kind":"event","title":"stepfun-ai/NextStep-1-f8ch16-Tokenizer","date":"2025-08-14T05:34:26+00:00","date_source":"source","source_url":"https://huggingface.co/stepfun-ai/NextStep-1-f8ch16-Tokenizer","signal_url":"https://onlylabs.fyi/signals/a5af86e7-03d5-4d99-a20c-7ccf3246e5d8","signal_json_url":"https://onlylabs.fyi/signals/a5af86e7-03d5-4d99-a20c-7ccf3246e5d8/signal.json","text":"model_released · stepfun-ai/NextStep-1-f8ch16-Tokenizer · signal_desk=releases · occurred_at=2025-08-14T05:34:26+00:00 · url=https://huggingface.co/stepfun-ai/NextStep-1-f8ch16-Tokenizer · hf_downloads=53 · hf_likes=15 · license=apache-2.0"},{"ref":"E25","kind":"event","title":"stepfun-ai/NextStep-1.1-Pretrain-256px","date":"2026-02-16T06:26:39+00:00","date_source":"source","source_url":"https://huggingface.co/stepfun-ai/NextStep-1.1-Pretrain-256px","signal_url":"https://onlylabs.fyi/signals/59a33bc4-341e-4c3e-8230-749ae55f0530","signal_json_url":"https://onlylabs.fyi/signals/59a33bc4-341e-4c3e-8230-749ae55f0530/signal.json","text":"model_released · stepfun-ai/NextStep-1.1-Pretrain-256px · signal_desk=releases · occurred_at=2026-02-16T06:26:39+00:00 · url=https://huggingface.co/stepfun-ai/NextStep-1.1-Pretrain-256px · hf_downloads=19 · hf_likes=13 · hf_params=14953814080 · pipeline=text-to-image · license=apache-2.0"},{"ref":"E26","kind":"event","title":"stepfun-ai/Qwen2.5-32B-DialogueReason","date":"2025-05-09T10:07:07+00:00","date_source":"source","source_url":"https://huggingface.co/stepfun-ai/Qwen2.5-32B-DialogueReason","signal_url":"https://onlylabs.fyi/signals/c9398311-d900-468c-a9b2-38535881d505","signal_json_url":"https://onlylabs.fyi/signals/c9398311-d900-468c-a9b2-38535881d505/signal.json","text":"model_released · stepfun-ai/Qwen2.5-32B-DialogueReason · signal_desk=releases · occurred_at=2025-05-09T10:07:07+00:00 · url=https://huggingface.co/stepfun-ai/Qwen2.5-32B-DialogueReason · hf_downloads=9 · hf_likes=12 · hf_params=32763876352 · license=apache-2.0"},{"ref":"E27","kind":"event","title":"stepfun-ai/PaCoRe","date":"2025-12-09T10:43:09+00:00","date_source":"source","source_url":"https://github.com/stepfun-ai/PaCoRe","signal_url":"https://onlylabs.fyi/signals/2c7576cb-3c36-45be-ba42-2215fa06fc21","signal_json_url":"https://onlylabs.fyi/signals/2c7576cb-3c36-45be-ba42-2215fa06fc21/signal.json","text":"repo_new · stepfun-ai/PaCoRe · signal_desk=repos · occurred_at=2025-12-09T10:43:09+00:00 · url=https://github.com/stepfun-ai/PaCoRe · stars=334 · hn=2 points/0 comments · raw={\"repo\":\"stepfun-ai/PaCoRe\",\"description\":\"PaCoRe: Learning to Scale Test-Time Compute with Parallel Coordinated Reasoning\",\"language\":\"Python\"}"},{"ref":"E28","kind":"event","title":"stepfun-ai/Step1X-Edit","date":"2025-04-23T09:53:08+00:00","date_source":"source","source_url":"https://github.com/stepfun-ai/Step1X-Edit","signal_url":"https://onlylabs.fyi/signals/5104756e-8899-44fd-a491-7b7a248f960d","signal_json_url":"https://onlylabs.fyi/signals/5104756e-8899-44fd-a491-7b7a248f960d/signal.json","text":"repo_new · stepfun-ai/Step1X-Edit · signal_desk=repos · occurred_at=2025-04-23T09:53:08+00:00 · url=https://github.com/stepfun-ai/Step1X-Edit · stars=2224 · raw={\"repo\":\"stepfun-ai/Step1X-Edit\",\"description\":\"A SOTA open-source image editing model, which aims to provide comparable performance against the closed-source models like GPT-4o and Gemini 2 Flash.\",\"language\":\"Python\"}"},{"ref":"E29","kind":"event","title":"stepfun-ai/StepFun-Prover-Preview-32B","date":"2025-08-13T08:26:04+00:00","date_source":"source","source_url":"https://huggingface.co/stepfun-ai/StepFun-Prover-Preview-32B","signal_url":"https://onlylabs.fyi/signals/2dca6d3c-dab8-4873-a520-529db3600124","signal_json_url":"https://onlylabs.fyi/signals/2dca6d3c-dab8-4873-a520-529db3600124/signal.json","text":"model_released · stepfun-ai/StepFun-Prover-Preview-32B · signal_desk=releases · occurred_at=2025-08-13T08:26:04+00:00 · url=https://huggingface.co/stepfun-ai/StepFun-Prover-Preview-32B · hf_downloads=60 · hf_likes=11 · hf_params=32763876352 · pipeline=text-generation · license=apache-2.0 · raw={\"derived_reason\":\"first-party-finetune\"}"},{"ref":"E30","kind":"event","title":"stepfun-ai/gelab-zero","date":"2025-11-28T14:42:44+00:00","date_source":"source","source_url":"https://github.com/stepfun-ai/gelab-zero","signal_url":"https://onlylabs.fyi/signals/1fcc2e72-93a3-49c5-b40a-1fa8f446dc3a","signal_json_url":"https://onlylabs.fyi/signals/1fcc2e72-93a3-49c5-b40a-1fa8f446dc3a/signal.json","text":"repo_new · stepfun-ai/gelab-zero · signal_desk=repos · occurred_at=2025-11-28T14:42:44+00:00 · url=https://github.com/stepfun-ai/gelab-zero · stars=2191 · raw={\"repo\":\"stepfun-ai/gelab-zero\",\"description\":\"STEP-GUI: The top GUI agent solution in the galaxy.  Developed by the StepFun-GELab team and powered by StepFun’s cutting-edge research capabilities.\",\"language\":\"Python\"}"},{"ref":"E31","kind":"event","title":"stepfun-ai/Step-3.5-Flash","date":"2026-01-31T02:57:03+00:00","date_source":"source","source_url":"https://github.com/stepfun-ai/Step-3.5-Flash","signal_url":"https://onlylabs.fyi/signals/b9c56539-05ed-4369-80f6-d40034c03d00","signal_json_url":"https://onlylabs.fyi/signals/b9c56539-05ed-4369-80f6-d40034c03d00/signal.json","text":"repo_new · stepfun-ai/Step-3.5-Flash · signal_desk=repos · occurred_at=2026-01-31T02:57:03+00:00 · url=https://github.com/stepfun-ai/Step-3.5-Flash · stars=2076 · raw={\"repo\":\"stepfun-ai/Step-3.5-Flash\",\"description\":\"Fast, Sharp & Reliable Agentic Intelligence\",\"language\":\"C++\"}"},{"ref":"E32","kind":"event","title":"stepfun-ai/StepFun-Formalizer-32B","date":"2025-08-19T02:59:21+00:00","date_source":"source","source_url":"https://huggingface.co/stepfun-ai/StepFun-Formalizer-32B","signal_url":"https://onlylabs.fyi/signals/cd1a3a1b-00f6-42e9-ac41-bef8788bae91","signal_json_url":"https://onlylabs.fyi/signals/cd1a3a1b-00f6-42e9-ac41-bef8788bae91/signal.json","text":"model_released · stepfun-ai/StepFun-Formalizer-32B · signal_desk=releases · occurred_at=2025-08-19T02:59:21+00:00 · url=https://huggingface.co/stepfun-ai/StepFun-Formalizer-32B · hf_downloads=682 · hf_likes=10 · hf_params=32763876352 · pipeline=text-generation · license=apache-2.0 · raw={\"derived_reason\":\"first-party-finetune\"}"},{"ref":"E33","kind":"event","title":"stepfun-ai/Step3-VL-10B-FP8","date":"2026-02-02T06:58:16+00:00","date_source":"source","source_url":"https://huggingface.co/stepfun-ai/Step3-VL-10B-FP8","signal_url":"https://onlylabs.fyi/signals/8d249566-8324-459c-a4ba-999588e66894","signal_json_url":"https://onlylabs.fyi/signals/8d249566-8324-459c-a4ba-999588e66894/signal.json","text":"model_released · stepfun-ai/Step3-VL-10B-FP8 · signal_desk=releases · occurred_at=2026-02-02T06:58:16+00:00 · url=https://huggingface.co/stepfun-ai/Step3-VL-10B-FP8 · hf_downloads=311 · hf_likes=10 · pipeline=image-text-to-text · license=apache-2.0 · raw={\"derived_reason\":\"first-party-finetune\"}"},{"ref":"E34","kind":"event","title":"stepfun-ai/RLVR-8B-0926","date":"2025-12-09T07:55:37+00:00","date_source":"source","source_url":"https://huggingface.co/stepfun-ai/RLVR-8B-0926","signal_url":"https://onlylabs.fyi/signals/b78b5531-8ce8-4809-93b2-37d6e84dd413","signal_json_url":"https://onlylabs.fyi/signals/b78b5531-8ce8-4809-93b2-37d6e84dd413/signal.json","text":"model_released · stepfun-ai/RLVR-8B-0926 · signal_desk=releases · occurred_at=2025-12-09T07:55:37+00:00 · url=https://huggingface.co/stepfun-ai/RLVR-8B-0926 · hf_downloads=50 · hf_likes=8 · hf_params=8190735360 · pipeline=text-generation · license=mit · raw={\"derived_reason\":\"first-party-finetune\"}"},{"ref":"E35","kind":"event","title":"stepfun-ai/SteptronOss","date":"2025-12-18T11:58:05+00:00","date_source":"source","source_url":"https://github.com/stepfun-ai/SteptronOss","signal_url":"https://onlylabs.fyi/signals/30426ad7-e4b4-49f6-85f5-fb4194ed460a","signal_json_url":"https://onlylabs.fyi/signals/30426ad7-e4b4-49f6-85f5-fb4194ed460a/signal.json","text":"repo_new · stepfun-ai/SteptronOss · signal_desk=repos · occurred_at=2025-12-18T11:58:05+00:00 · url=https://github.com/stepfun-ai/SteptronOss · stars=575 · hn=1 points/0 comments · raw={\"repo\":\"stepfun-ai/SteptronOss\",\"description\":\"A lightweight, AI-native training framework for large language models. Designed for fast iteration, reproducible experiments, and modular configuration across SFT, RLVR, and evaluation workflows.\",\"language\":\"Python\"}"},{"ref":"E36","kind":"event","title":"stepfun-ai/Step-Audio2","date":"2025-07-15T09:14:32+00:00","date_source":"source","source_url":"https://github.com/stepfun-ai/Step-Audio2","signal_url":"https://onlylabs.fyi/signals/736cc01e-0975-4c81-8e42-a41217c2cf38","signal_json_url":"https://onlylabs.fyi/signals/736cc01e-0975-4c81-8e42-a41217c2cf38/signal.json","text":"repo_new · stepfun-ai/Step-Audio2 · signal_desk=repos · occurred_at=2025-07-15T09:14:32+00:00 · url=https://github.com/stepfun-ai/Step-Audio2 · stars=1460 · raw={\"repo\":\"stepfun-ai/Step-Audio2\",\"description\":\"Step-Audio 2 is an end-to-end multi-modal large language model designed for industry-strength audio understanding and speech conversation.\",\"language\":\"Python\"}"},{"ref":"E37","kind":"event","title":"stepfun-ai/NextStep-1.1-Pretrain","date":"2025-12-24T03:11:30+00:00","date_source":"source","source_url":"https://huggingface.co/stepfun-ai/NextStep-1.1-Pretrain","signal_url":"https://onlylabs.fyi/signals/f6b2bd36-f3d7-4bc2-bb3f-77d53e812199","signal_json_url":"https://onlylabs.fyi/signals/f6b2bd36-f3d7-4bc2-bb3f-77d53e812199/signal.json","text":"model_released · stepfun-ai/NextStep-1.1-Pretrain · signal_desk=releases · occurred_at=2025-12-24T03:11:30+00:00 · url=https://huggingface.co/stepfun-ai/NextStep-1.1-Pretrain · hf_downloads=13 · hf_likes=7 · hf_params=14953814080 · pipeline=text-to-image · license=apache-2.0"},{"ref":"E38","kind":"event","title":"stepfun-ai/StepFun-Formalizer-7B","date":"2025-08-19T02:27:17+00:00","date_source":"source","source_url":"https://huggingface.co/stepfun-ai/StepFun-Formalizer-7B","signal_url":"https://onlylabs.fyi/signals/19b4f3dd-1626-42c9-9fac-6f4a10820ee0","signal_json_url":"https://onlylabs.fyi/signals/19b4f3dd-1626-42c9-9fac-6f4a10820ee0/signal.json","text":"model_released · stepfun-ai/StepFun-Formalizer-7B · signal_desk=releases · occurred_at=2025-08-19T02:27:17+00:00 · url=https://huggingface.co/stepfun-ai/StepFun-Formalizer-7B · hf_downloads=472 · hf_likes=6 · hf_params=7615616512 · pipeline=text-generation · license=apache-2.0 · raw={\"derived_reason\":\"first-party-finetune\"}"},{"ref":"E39","kind":"event","title":"stepfun-ai/Step1X-Edit-v1p1-diffusers","date":"2025-08-27T07:32:40+00:00","date_source":"source","source_url":"https://huggingface.co/stepfun-ai/Step1X-Edit-v1p1-diffusers","signal_url":"https://onlylabs.fyi/signals/57957373-f56c-481c-a849-3727564bcdd1","signal_json_url":"https://onlylabs.fyi/signals/57957373-f56c-481c-a849-3727564bcdd1/signal.json","text":"model_released · stepfun-ai/Step1X-Edit-v1p1-diffusers · signal_desk=releases · occurred_at=2025-08-27T07:32:40+00:00 · url=https://huggingface.co/stepfun-ai/Step1X-Edit-v1p1-diffusers · hf_downloads=46 · hf_likes=5 · pipeline=image-to-image · license=apache-2.0"},{"ref":"E40","kind":"event","title":"stepfun-ai/Step-Audio-EditX","date":"2025-10-29T11:54:17+00:00","date_source":"source","source_url":"https://github.com/stepfun-ai/Step-Audio-EditX","signal_url":"https://onlylabs.fyi/signals/8fe7a844-0db1-4144-8532-c2a931682681","signal_json_url":"https://onlylabs.fyi/signals/8fe7a844-0db1-4144-8532-c2a931682681/signal.json","text":"repo_new · stepfun-ai/Step-Audio-EditX · signal_desk=repos · occurred_at=2025-10-29T11:54:17+00:00 · url=https://github.com/stepfun-ai/Step-Audio-EditX · stars=929 · raw={\"repo\":\"stepfun-ai/Step-Audio-EditX\",\"description\":\"A powerful 3B-parameter, LLM-based Reinforcement Learning audio edit model excels at editing emotion, speaking style, and paralinguistics, and features robust zero-shot text-to-speech\",\"language\":\"Python\"}"},{"ref":"E41","kind":"event","title":"stepfun-ai/Step1X-3D","date":"2025-05-13T03:42:54+00:00","date_source":"source","source_url":"https://github.com/stepfun-ai/Step1X-3D","signal_url":"https://onlylabs.fyi/signals/3040d01f-41dd-4cad-b8a5-59df34df8123","signal_json_url":"https://onlylabs.fyi/signals/3040d01f-41dd-4cad-b8a5-59df34df8123/signal.json","text":"repo_new · stepfun-ai/Step1X-3D · signal_desk=repos · occurred_at=2025-05-13T03:42:54+00:00 · url=https://github.com/stepfun-ai/Step1X-3D · stars=869 · raw={\"repo\":\"stepfun-ai/Step1X-3D\",\"description\":\"Step1X-3D: Towards High-Fidelity and Controllable Generation of Textured 3D Assets\",\"language\":\"Python\"}"},{"ref":"E42","kind":"event","title":"stepfun-ai/NextStep-1","date":"2025-08-14T08:50:25+00:00","date_source":"source","source_url":"https://github.com/stepfun-ai/NextStep-1","signal_url":"https://onlylabs.fyi/signals/1ef8773f-1e64-4e8b-ba26-e9db4251de8b","signal_json_url":"https://onlylabs.fyi/signals/1ef8773f-1e64-4e8b-ba26-e9db4251de8b/signal.json","text":"repo_new · stepfun-ai/NextStep-1 · signal_desk=repos · occurred_at=2025-08-14T08:50:25+00:00 · url=https://github.com/stepfun-ai/NextStep-1 · stars=689 · raw={\"repo\":\"stepfun-ai/NextStep-1\",\"description\":\"[🚀 ICLR 2026 Oral] NextStep-1: SOTA Autogressive Image Generation with Continuous Tokens. A research project developed by the StepFun’s Multimodal Intelligence team.\",\"language\":\"Python\"}"},{"ref":"E43","kind":"event","title":"stepfun-ai/Step-Audio-R1","date":"2025-11-11T02:38:13+00:00","date_source":"source","source_url":"https://github.com/stepfun-ai/Step-Audio-R1","signal_url":"https://onlylabs.fyi/signals/76b4d95e-e479-4749-92eb-c272aaa58c8a","signal_json_url":"https://onlylabs.fyi/signals/76b4d95e-e479-4749-92eb-c272aaa58c8a/signal.json","text":"repo_new · stepfun-ai/Step-Audio-R1 · signal_desk=repos · occurred_at=2025-11-11T02:38:13+00:00 · url=https://github.com/stepfun-ai/Step-Audio-R1 · stars=673 · raw={\"repo\":\"stepfun-ai/Step-Audio-R1\",\"language\":\"Python\"}"},{"ref":"E44","kind":"event","title":"stepfun-ai/StepFun-Prover-Preview-7B","date":"2025-08-13T08:10:19+00:00","date_source":"source","source_url":"https://huggingface.co/stepfun-ai/StepFun-Prover-Preview-7B","signal_url":"https://onlylabs.fyi/signals/cb9aede2-4f2c-44f1-a371-286e63a1a0ef","signal_json_url":"https://onlylabs.fyi/signals/cb9aede2-4f2c-44f1-a371-286e63a1a0ef/signal.json","text":"model_released · stepfun-ai/StepFun-Prover-Preview-7B · signal_desk=releases · occurred_at=2025-08-13T08:10:19+00:00 · url=https://huggingface.co/stepfun-ai/StepFun-Prover-Preview-7B · hf_downloads=70 · hf_likes=3 · hf_params=7615616512 · pipeline=text-generation · license=apache-2.0 · raw={\"derived_reason\":\"first-party-finetune\"}"},{"ref":"E45","kind":"event","title":"stepfun-ai/StepDeepResearch","date":"2025-11-26T06:02:31+00:00","date_source":"source","source_url":"https://github.com/stepfun-ai/StepDeepResearch","signal_url":"https://onlylabs.fyi/signals/f28c49ee-84ff-4a3c-b4ff-cdb2b59b1363","signal_json_url":"https://onlylabs.fyi/signals/f28c49ee-84ff-4a3c-b4ff-cdb2b59b1363/signal.json","text":"repo_new · stepfun-ai/StepDeepResearch · signal_desk=repos · occurred_at=2025-11-26T06:02:31+00:00 · url=https://github.com/stepfun-ai/StepDeepResearch · stars=561 · raw={\"repo\":\"stepfun-ai/StepDeepResearch\",\"description\":\"Step-DeepResearch\",\"language\":\"Python\"}"},{"ref":"E46","kind":"event","title":"stepfun-ai/Step3","date":"2025-07-25T03:11:20+00:00","date_source":"source","source_url":"https://github.com/stepfun-ai/Step3","signal_url":"https://onlylabs.fyi/signals/771e0320-2002-4f2e-8b8e-b6ecfd0bd520","signal_json_url":"https://onlylabs.fyi/signals/771e0320-2002-4f2e-8b8e-b6ecfd0bd520/signal.json","text":"repo_new · stepfun-ai/Step3 · signal_desk=repos · occurred_at=2025-07-25T03:11:20+00:00 · url=https://github.com/stepfun-ai/Step3 · stars=453 · raw={\"repo\":\"stepfun-ai/Step3\"}"},{"ref":"E47","kind":"event","title":"stepfun-ai/Step3-VL-10B","date":"2026-01-13T09:13:12+00:00","date_source":"source","source_url":"https://github.com/stepfun-ai/Step3-VL-10B","signal_url":"https://onlylabs.fyi/signals/d57b4ea8-177f-4362-b4a3-18a1f1e3f575","signal_json_url":"https://onlylabs.fyi/signals/d57b4ea8-177f-4362-b4a3-18a1f1e3f575/signal.json","text":"repo_new · stepfun-ai/Step3-VL-10B · signal_desk=repos · occurred_at=2026-01-13T09:13:12+00:00 · url=https://github.com/stepfun-ai/Step3-VL-10B · stars=407 · raw={\"repo\":\"stepfun-ai/Step3-VL-10B\",\"description\":\"Step3-VL-10B: A compact yet frontier multimodal model achieving SOTA performance at the 10B scale, matching open-source models 10-20x its size.\"}"},{"ref":"E48","kind":"event","title":"stepfun-ai/Step-Video-TI2V","date":"2025-03-06T04:42:11+00:00","date_source":"source","source_url":"https://github.com/stepfun-ai/Step-Video-TI2V","signal_url":"https://onlylabs.fyi/signals/33465a53-c301-4e1c-95bc-48f0068c0657","signal_json_url":"https://onlylabs.fyi/signals/33465a53-c301-4e1c-95bc-48f0068c0657/signal.json","text":"repo_new · stepfun-ai/Step-Video-TI2V · signal_desk=repos · occurred_at=2025-03-06T04:42:11+00:00 · url=https://github.com/stepfun-ai/Step-Video-TI2V · stars=375 · raw={\"repo\":\"stepfun-ai/Step-Video-TI2V\",\"language\":\"Python\"}"},{"ref":"E49","kind":"event","title":"stepfun-ai/StepMesh","date":"2025-07-21T11:47:08+00:00","date_source":"source","source_url":"https://github.com/stepfun-ai/StepMesh","signal_url":"https://onlylabs.fyi/signals/b55dcb30-239e-47b6-aa69-9625dc783d43","signal_json_url":"https://onlylabs.fyi/signals/b55dcb30-239e-47b6-aa69-9625dc783d43/signal.json","text":"repo_new · stepfun-ai/StepMesh · signal_desk=repos · occurred_at=2025-07-21T11:47:08+00:00 · url=https://github.com/stepfun-ai/StepMesh · stars=367 · raw={\"repo\":\"stepfun-ai/StepMesh\",\"language\":\"C++\"}"},{"ref":"E50","kind":"event","title":"stepfun-ai/llama.cpp","date":"2026-03-16T07:20:37+00:00","date_source":"source","source_url":"https://github.com/stepfun-ai/llama.cpp","signal_url":"https://onlylabs.fyi/signals/411e46af-07c9-427d-8522-f0502d830c66","signal_json_url":"https://onlylabs.fyi/signals/411e46af-07c9-427d-8522-f0502d830c66/signal.json","text":"repo_forked · stepfun-ai/llama.cpp · signal_desk=forks · occurred_at=2026-03-16T07:20:37+00:00 · url=https://github.com/stepfun-ai/llama.cpp · stars=5 · raw={\"repo\":\"stepfun-ai/llama.cpp\",\"parent\":\"ggml-org/llama.cpp\"}"},{"ref":"E51","kind":"event","title":"stepfun-ai/Step-Realtime-Console","date":"2025-03-12T10:09:47+00:00","date_source":"source","source_url":"https://github.com/stepfun-ai/Step-Realtime-Console","signal_url":"https://onlylabs.fyi/signals/2b610a31-60bf-40d5-bcf8-27ec02be707c","signal_json_url":"https://onlylabs.fyi/signals/2b610a31-60bf-40d5-bcf8-27ec02be707c/signal.json","text":"repo_new · stepfun-ai/Step-Realtime-Console · signal_desk=repos · occurred_at=2025-03-12T10:09:47+00:00 · url=https://github.com/stepfun-ai/Step-Realtime-Console · stars=74 · raw={\"repo\":\"stepfun-ai/Step-Realtime-Console\",\"description\":\"Step-Realtime-Console\",\"language\":\"JavaScript\"}"},{"ref":"E52","kind":"event","title":"stepfun-ai/GEBench","date":"2026-02-09T13:18:49+00:00","date_source":"source","source_url":"https://github.com/stepfun-ai/GEBench","signal_url":"https://onlylabs.fyi/signals/34766301-c7b4-4b29-a48f-7449c66ed4f5","signal_json_url":"https://onlylabs.fyi/signals/34766301-c7b4-4b29-a48f-7449c66ed4f5/signal.json","text":"repo_new · stepfun-ai/GEBench · signal_desk=repos · occurred_at=2026-02-09T13:18:49+00:00 · url=https://github.com/stepfun-ai/GEBench · stars=54 · raw={\"repo\":\"stepfun-ai/GEBench\",\"language\":\"Python\"}"},{"ref":"E53","kind":"event","title":"stepfun-ai/vllm","date":"2025-02-18T07:53:03+00:00","date_source":"source","source_url":"https://github.com/stepfun-ai/vllm","signal_url":"https://onlylabs.fyi/signals/a14be8d5-87fe-47a5-8e4a-1ace1cebce42","signal_json_url":"https://onlylabs.fyi/signals/a14be8d5-87fe-47a5-8e4a-1ace1cebce42/signal.json","text":"repo_forked · stepfun-ai/vllm · signal_desk=forks · occurred_at=2025-02-18T07:53:03+00:00 · url=https://github.com/stepfun-ai/vllm · stars=44 · raw={\"repo\":\"stepfun-ai/vllm\",\"parent\":\"vllm-project/vllm\"}"},{"ref":"E54","kind":"event","title":"stepfun-ai/ComfyUI-StepVideo","date":"2025-03-11T04:12:38+00:00","date_source":"source","source_url":"https://github.com/stepfun-ai/ComfyUI-StepVideo","signal_url":"https://onlylabs.fyi/signals/f476acb7-c83c-410c-bd65-1631d0b3d4c5","signal_json_url":"https://onlylabs.fyi/signals/f476acb7-c83c-410c-bd65-1631d0b3d4c5/signal.json","text":"repo_new · stepfun-ai/ComfyUI-StepVideo · signal_desk=repos · occurred_at=2025-03-11T04:12:38+00:00 · url=https://github.com/stepfun-ai/ComfyUI-StepVideo · stars=43 · raw={\"repo\":\"stepfun-ai/ComfyUI-StepVideo\",\"language\":\"Python\"}"},{"ref":"E55","kind":"event","title":"stepfun-ai/StepFun-Prover-Preview","date":"2025-08-13T09:19:23+00:00","date_source":"source","source_url":"https://github.com/stepfun-ai/StepFun-Prover-Preview","signal_url":"https://onlylabs.fyi/signals/6a4a16e8-4c30-41b3-9a8c-8189e613d215","signal_json_url":"https://onlylabs.fyi/signals/6a4a16e8-4c30-41b3-9a8c-8189e613d215/signal.json","text":"repo_new · stepfun-ai/StepFun-Prover-Preview · signal_desk=repos · occurred_at=2025-08-13T09:19:23+00:00 · url=https://github.com/stepfun-ai/StepFun-Prover-Preview · stars=35 · raw={\"repo\":\"stepfun-ai/StepFun-Prover-Preview\",\"description\":\"Large language models designed for formal theorem proving through tool-integrated reasoning.\"}"},{"ref":"E56","kind":"event","title":"stepfun-ai/StepFun-Formalizer","date":"2025-08-19T03:34:06+00:00","date_source":"source","source_url":"https://github.com/stepfun-ai/StepFun-Formalizer","signal_url":"https://onlylabs.fyi/signals/9fae1ecf-bfbd-4f4d-b28b-ecf0bb3bfacc","signal_json_url":"https://onlylabs.fyi/signals/9fae1ecf-bfbd-4f4d-b28b-ecf0bb3bfacc/signal.json","text":"repo_new · stepfun-ai/StepFun-Formalizer · signal_desk=repos · occurred_at=2025-08-19T03:34:06+00:00 · url=https://github.com/stepfun-ai/StepFun-Formalizer · stars=30 · raw={\"repo\":\"stepfun-ai/StepFun-Formalizer\",\"description\":\"StepFun-Formalizer: Unlocking the Autoformalization Potential of LLMs through Knowledge-Reasoning Fusion\",\"language\":\"Python\"}"},{"ref":"E57","kind":"event","title":"stepfun-ai/Step-Audio","date":"2025-02-11T05:35:12+00:00","date_source":"source","source_url":"https://github.com/stepfun-ai/Step-Audio","signal_url":"https://onlylabs.fyi/signals/a0929a83-8ca3-4979-ae63-39878e29e4a9","signal_json_url":"https://onlylabs.fyi/signals/a0929a83-8ca3-4979-ae63-39878e29e4a9/signal.json","text":"repo_new · stepfun-ai/Step-Audio · signal_desk=repos · occurred_at=2025-02-11T05:35:12+00:00 · url=https://github.com/stepfun-ai/Step-Audio · stars=27 · raw={\"repo\":\"stepfun-ai/Step-Audio\",\"language\":\"Python\"}"},{"ref":"E58","kind":"event","title":"stepfun-ai/StepAudio-Skills","date":"2026-03-09T12:48:01+00:00","date_source":"source","source_url":"https://github.com/stepfun-ai/StepAudio-Skills","signal_url":"https://onlylabs.fyi/signals/b9789e37-6115-4fc1-aec9-68b73e2d2186","signal_json_url":"https://onlylabs.fyi/signals/b9789e37-6115-4fc1-aec9-68b73e2d2186/signal.json","text":"repo_new · stepfun-ai/StepAudio-Skills · signal_desk=repos · occurred_at=2026-03-09T12:48:01+00:00 · url=https://github.com/stepfun-ai/StepAudio-Skills · stars=26 · raw={\"repo\":\"stepfun-ai/StepAudio-Skills\",\"description\":\"Audio skills for Claw\",\"language\":\"Python\"}"},{"ref":"E59","kind":"event","title":"stepfun-ai/Step-Audio-Edit-Benchmark","date":"2025-11-20T06:36:16+00:00","date_source":"source","source_url":"https://github.com/stepfun-ai/Step-Audio-Edit-Benchmark","signal_url":"https://onlylabs.fyi/signals/d675a0c1-be4c-4c47-be18-84e488dea913","signal_json_url":"https://onlylabs.fyi/signals/d675a0c1-be4c-4c47-be18-84e488dea913/signal.json","text":"repo_new · stepfun-ai/Step-Audio-Edit-Benchmark · signal_desk=repos · occurred_at=2025-11-20T06:36:16+00:00 · url=https://github.com/stepfun-ai/Step-Audio-Edit-Benchmark · stars=20 · raw={\"repo\":\"stepfun-ai/Step-Audio-Edit-Benchmark\",\"language\":\"Python\"}"},{"ref":"E60","kind":"event","title":"stepfun-ai/InfiniteHBD-Trace","date":"2025-05-21T02:16:57+00:00","date_source":"source","source_url":"https://github.com/stepfun-ai/InfiniteHBD-Trace","signal_url":"https://onlylabs.fyi/signals/13424389-1504-48dc-95c2-84c7244973bf","signal_json_url":"https://onlylabs.fyi/signals/13424389-1504-48dc-95c2-84c7244973bf/signal.json","text":"repo_new · stepfun-ai/InfiniteHBD-Trace · signal_desk=repos · occurred_at=2025-05-21T02:16:57+00:00 · url=https://github.com/stepfun-ai/InfiniteHBD-Trace · stars=18 · raw={\"repo\":\"stepfun-ai/InfiniteHBD-Trace\"}"}]}