A dense desk for emerging AI labs: what they are hiring for, releasing, forking, publishing, and accumulating as evidence for future analysis.
Upstage is executing a platform consolidation play, moving from a pure models-and-APIs company toward an integrated AI stack combining proprietary Solar LLMs, acquired portal (Daum) assets, and agentic platforms (Timely). Core thesis:…
AI21 Labs' recent signal set shows a lab pivoting from general-purpose frontier models toward enterprise-grade *specialized* intelligence and verification-first agent optimization. Three evidence clusters converge on this: (1) compliance…
Sarvam AI is at a commercialization inflection: it is simultaneously releasing open-weight frontier models — Sarvam 30B and 105B, speech (Saaras V4), and vision (Sarvam Vision 2.1) — while hiring most aggressively into go-to-market,…
InclusionAI (Ant Group's AGI arm) is running a two-track play: a frontier open-weights text line (Ling) now reaching a 560B/25B-active hybrid-linear MoE flagship, layered with fast multi-modal and vertical specialization — healthcare…
LG AI Research's EXAONE program is shifting from sequential research LLM releases toward a hybrid, applied, and industrialized portfolio. The evidence shows a multi-generation model family — dense 1.2B–32B EXAONE checkpoints, a…
Reka's public signal has decisively shifted from a general multimodal-LLM / enterprise-automation lab (Reka Flash/Core/Edge, Nexus) toward physical-world AI: 'omni world models' that unify language, images, video, and audio in a single…
Nous Research is running two visible tracks at once. One track keeps shipping open-weight model finetunes under permissive licenses — Hermes-4 (14B/70B/36B), NousCoder-14B, and nomos-1. The other track, which dominates the October 2026…
IBM is shipping an Apache-2.0 open-model portfolio - dense reasoning text models (Granite 4.2), multilingual speech (Granite Speech 5.0), modular single-endpoint models (Granite Switch 4.1), and time-series forecasting (Granite Timeseries)…
OpenBMB's evidence cluster resolves into two parallel thrusts. First, a research/engineering bet on on-device efficiency: small and MoE multimodal/language models, speculative decoding (DSpark), and a C++ omni inference engine…
Arcee AI has crossed from a model-tuning and merging shop into a first-party frontier open-weight lab. Its Trinity family runs from a 400B sparse Mixture-of-Experts model down to 6B dense, trained in-house with a claimed ~$20M budget, and…
Inception is a Palo Alto–based model lab built on a single architectural bet: replacing autoregressive, token-by-token decoding with diffusion (dLLMs) so that latency, not raw benchmark intelligence, becomes the differentiated product. Its…
StepFun is a Shanghai-based multimodal frontier lab — founded in 2023 and led by Daxin Jiang, an ex-Microsoft speech/Azure AI executive — that uses aggressive open-sourcing of weights, benchmarks, and training/inference tooling as its…
XiaomiMiMo is executing a full-stack, open-source AI strategy that spans text reasoning, vision, audio, embodied AI, and coding agents — with an accelerating pivot toward the Agent era in mid-2026. The lab's evidence trail reveals a…
{"content":"## Thesis\n\nMeituan's LongCat lab is a full-stack frontier AI organization: it ships frontier-scale MoE language models (LongCat-2.0 at 1.6T total / ~48B activated parameters) while simultaneously building the inference…
Neolab has 14 tracked organizations with 2,724 loaded public signals: 346 hiring, 146 forks, 1,064 releases or model cards, 695 talking, and 473 repos. Upstage (Solar) leads this desk with 468 loaded signals. Latest contextified signal: Upstage (Solar) - People 05. 14 standing agent analyses are generated for this category.
tracks 14 organizations
has loaded 2,724 public signals
has hiring signal count 346
has fork signal count 146
Inspect Neolab as the category-level intelligence desk with 14 tracked organizations and 2,724 loaded public signals. Rank priority accounts across hiring (346), forks (146), releases (1,064), talking (695), and repos (473). Identify account momentum, source gaps, and analysis-agent dispatch work for this category.
Upstage (Solar)Upstage (Solar)468 signals