# Labs: AI, data and research labs, full data

> Generated from https://labs.fru.dev. One block per lab. Facts from each lab's own pages, Wikidata, Wikipedia and the papers themselves, with the source linked; leaders, models and papers come from the fru.dev sites that track them.

## 01.AI
https://labs.fru.dev/labs/01-ai | Startup lab | founded 2023 | Beijing, China | https://www.01.ai
Known for: Kai-Fu Lee's lab behind the open Yi models, now focused on enterprise AI products.
Leaders: Kai-Fu Lee (Founder; https://www.01.ai/)
Flagship model: Yi-34B (https://huggingface.co/01-ai/Yi-34B)
- Landmark 2024: Yi: Open Foundation Models by 01.AI (https://arxiv.org/abs/2403.04652)
Links: https://github.com/01-ai https://huggingface.co/01-ai
Sources: https://www.wikidata.org/wiki/Q130692307 https://www.01.ai/

## Ai2 (Allen Institute for AI)
https://labs.fru.dev/labs/ai2 | Nonprofit | founded 2014 | Seattle, United States | https://allenai.org
Known for: Fully open models with their data and training code: OLMo, Tulu, Molmo, and the Dolma corpus.
Leaders: Peter Clark (Interim CEO; https://en.wikipedia.org/wiki/Allen_Institute_for_AI)
Flagship model: OLMo (https://allenai.org/olmo)
- Landmark 2018: Deep contextualized word representations (ELMo) (https://arxiv.org/abs/1802.05365)
- Landmark 2024: OLMo: Accelerating the Science of Language Models (https://arxiv.org/abs/2402.00838)
- Landmark 2024: Dolma: an Open Corpus of Three Trillion Tokens for Language Model Pretraining Research (https://arxiv.org/abs/2402.00159)
- Landmark 2024: Molmo and PixMo: Open Weights and Open Data for State-of-the-Art Vision-Language Models (https://arxiv.org/abs/2409.17146)
- Landmark 2024: Tulu 3: Pushing Frontiers in Open Language Model Post-Training (https://arxiv.org/abs/2411.15124)
Links: https://github.com/allenai https://huggingface.co/allenai https://x.com/allen_ai
Sources: https://www.wikidata.org/wiki/Q16002567 https://en.wikipedia.org/wiki/Allen_Institute_for_AI

## AI21 Labs
https://labs.fru.dev/labs/ai21-labs | Startup lab | founded 2017 | Tel Aviv, Israel | https://www.ai21.com
Known for: Jamba, the first production-scale hybrid of Transformer and Mamba layers.
Leaders: Ori Goshen (CEO and co-founder; https://www.ai21.com/about/)
- Landmark 2024: Jamba: A Hybrid Transformer-Mamba Language Model (https://arxiv.org/abs/2403.19887)
- Landmark 2024: Jamba-1.5: Hybrid Transformer-Mamba Models at Scale (https://arxiv.org/abs/2408.12570)
Links: https://github.com/AI21Labs https://huggingface.co/ai21labs https://x.com/AI21Labs
Sources: https://www.wikidata.org/wiki/Q113030551 https://www.ai21.com/about/

## Alibaba Qwen
https://labs.fru.dev/labs/qwen | Company lab of Alibaba | founded 2023 | Hangzhou, China | https://qwen.ai
Known for: The Qwen family, the most downloaded open-weight models, from Alibaba Cloud's Tongyi lab.
Flagship model: Qwen3.8 Max (0902) (https://models.fru.dev/models/qwen3-8-max)
- Landmark 2023: Qwen Technical Report (https://arxiv.org/abs/2309.16609)
- Landmark 2024: Qwen2.5 Technical Report (https://arxiv.org/abs/2412.15115)
- Landmark 2025: Qwen3 Technical Report (https://arxiv.org/abs/2505.09388)
Links: https://github.com/QwenLM https://huggingface.co/Qwen https://x.com/Alibaba_Qwen
Sources: https://www.wikidata.org/wiki/Q130234299

## Amazon AGI and Amazon Science
https://labs.fru.dev/labs/amazon-agi | Company lab of Amazon | Seattle, United States | https://www.amazon.science
Known for: The Amazon Nova models and Nova Act agents, plus Chronos forecasting models from Amazon Science.
Leaders: Peter DeSantis (Leads the combined AI, silicon and quantum computing organization; https://www.geekwire.com/2025/amazon-ai-chief-rohit-prasad-leaving-infrastructure-exec-peter-desantis-to-lead-unified-ai-group/)
Flagship model: Amazon Nova (https://aws.amazon.com/ai/generative-ai/nova/)
- Landmark 2024: Chronos: Learning the Language of Time Series (https://arxiv.org/abs/2403.07815)
- Landmark 2024: The Amazon Nova family of models: technical report and model card (https://www.amazon.science/publications/the-amazon-nova-family-of-models-technical-report-and-model-card)
Links: https://github.com/amazon-science https://huggingface.co/amazon https://x.com/AmazonScience
Sources: https://www.amazon.science/

## Anthropic
https://labs.fru.dev/labs/anthropic | Startup lab | founded 2021 | San Francisco, United States | https://www.anthropic.com
Known for: The Claude models, Constitutional AI and interpretability research on what happens inside models.
Leaders: Dario Amodei (Chief Executive Officer and co-founder; https://www.anthropic.com/company); Rahul Patil (Chief Technology Officer; https://techcrunch.com/2025/10/02/anthropic-hires-new-cto-with-focus-on-ai-infrastructure/); Jared Kaplan (Chief Science Officer and co-founder; https://en.wikipedia.org/wiki/Jared_Kaplan)
Flagship model: Claude Fable 5.1 (https://models.fru.dev/models/claude-fable-5-1)
- Landmark 2022: Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback (https://arxiv.org/abs/2204.05862)
- Landmark 2022: Constitutional AI: Harmlessness from AI Feedback (https://arxiv.org/abs/2212.08073)
- Landmark 2024: Scaling Monosemanticity: Extracting Interpretable Features from Claude 3 Sonnet (https://transformer-circuits.pub/2024/scaling-monosemanticity/)
- Landmark 2024: Introducing the Model Context Protocol (https://www.anthropic.com/news/model-context-protocol)
- Funding 2026-05-28: Series H $65B, at a $965B post-money valuation, led by Altimeter, Dragoneer, Greenoaks and Sequoia (https://www.anthropic.com/news/series-h)
- Funding 2026-02-12: Series G $30B, at a $380B post-money valuation, led by GIC and Coatue (https://www.anthropic.com/news/anthropic-raises-30-billion-series-g-funding-380-billion-post-money-valuation)
- Compute 2025-10-23: Google Cloud, Up to one million TPUs, well over a gigawatt online in 2026 (https://www.anthropic.com/news/expanding-our-use-of-google-cloud-tpus-and-services)
- Compute 2025-11-18: Microsoft and NVIDIA, $30B of Azure compute, up to one gigawatt more (https://www.anthropic.com/news/microsoft-nvidia-anthropic-announce-strategic-partnerships)
- Compute 2026-04-06: Google and Broadcom, Multiple gigawatts of next-generation TPU capacity from 2027 (https://www.anthropic.com/news/google-broadcom-partnership-compute)
- Compute 2026-04-20: Amazon, Up to 5 GW of Trainium capacity (https://www.anthropic.com/news/anthropic-amazon-compute)
- Compute 2026-05-06: SpaceX, Compute partnership (https://www.anthropic.com/news/higher-limits-spacex)
Links: https://github.com/anthropics https://huggingface.co/anthropic https://x.com/AnthropicAI
Sources: https://www.wikidata.org/wiki/Q116758847

## Apple Machine Learning Research
https://labs.fru.dev/labs/apple-ml-research | Company lab of Apple | Cupertino, United States | https://machinelearning.apple.com
Known for: On-device models for Apple Intelligence, the MLX framework for Apple silicon, and studies of reasoning limits.
Leaders: Amar Subramanya (vice president of AI; https://www.apple.com/newsroom/2025/12/john-giannandrea-to-retire-from-apple/)
- Landmark 2024: OpenELM: An Efficient Language Model Family with Open Training and Inference Framework (https://arxiv.org/abs/2404.14619)
- Landmark 2023: MLX: an array framework for Apple silicon (https://github.com/ml-explore/mlx)
- Landmark 2024: Apple Intelligence Foundation Language Models (https://arxiv.org/abs/2407.21075)
- Landmark 2025: The Illusion of Thinking (https://machinelearning.apple.com/research/illusion-of-thinking)
Links: https://github.com/apple https://huggingface.co/apple
Sources: https://machinelearning.apple.com/

## BAAI (Beijing Academy of AI)
https://labs.fru.dev/labs/baai | Nonprofit | founded 2018 | Beijing, China | https://www.baai.ac.cn/en/
Known for: The BGE embedding models used across open retrieval stacks, the Emu multimodal models and FlagOpen.
Flagship model: BGE-M3 (https://huggingface.co/BAAI/bge-m3)
- Landmark 2023: C-Pack: Packed Resources For General Chinese Embeddings (BGE) (https://arxiv.org/abs/2309.07597)
- Landmark 2024: Emu3: Next-Token Prediction is All You Need (https://arxiv.org/abs/2409.18869)
Links: https://github.com/FlagOpen https://huggingface.co/BAAI
Sources: https://www.wikidata.org/wiki/Q107518033 https://en.wikipedia.org/wiki/Beijing_Academy_of_Artificial_Intelligence

## Baidu (ERNIE)
https://labs.fru.dev/labs/baidu-ernie | Company lab of Baidu | Beijing, China | https://research.baidu.com
Known for: The ERNIE models and PaddlePaddle, and earlier Deep Speech 2 from its Silicon Valley lab.
Leaders: Robin Li (Chief Executive Officer; http://ir.baidu.com/phoenix.zhtml?c=188488&p=irol-govmanage)
Flagship model: Ernie 5.1 (https://models.fru.dev/models/ernie-5-1)
- Landmark 2015: Deep Speech 2: End-to-End Speech Recognition in English and Mandarin (https://arxiv.org/abs/1512.02595)
- Landmark 2019: ERNIE: Enhanced Representation through Knowledge Integration (https://arxiv.org/abs/1904.09223)
Links: https://github.com/PaddlePaddle https://huggingface.co/baidu
Sources: https://www.wikidata.org/wiki/Q14772 https://research.baidu.com/

## Berkeley AI Research (BAIR)
https://labs.fru.dev/labs/berkeley-bair | Academic of UC Berkeley | Berkeley, California, United States | https://bair.berkeley.edu
Known for: Diffusion models (DDPM), NeRF, robot learning, and open chat models such as Vicuna.
- Landmark 2020: Denoising Diffusion Probabilistic Models (https://arxiv.org/abs/2006.11239)
- Landmark 2020: NeRF: Representing Scenes as Neural Radiance Fields for View Synthesis (https://arxiv.org/abs/2003.08934)
- Landmark 2023: Vicuna: An Open-Source Chatbot Impressing GPT-4 with 90% ChatGPT Quality (https://lmsys.org/blog/2023-03-30-vicuna/)
- Landmark 2023: Gorilla: Large Language Model Connected with Massive APIs (https://arxiv.org/abs/2305.15334)
Links: https://x.com/berkeley_ai
Sources: https://bair.berkeley.edu/about

## Berkeley Sky Computing Lab
https://labs.fru.dev/labs/berkeley-sky | Academic of UC Berkeley | Berkeley, California, United States | https://sky.cs.berkeley.edu
Known for: The systems lab after AMPLab (Spark) and RISELab (Ray): vLLM, SkyPilot and Chatbot Arena.
Leaders: Ion Stoica (Core faculty; https://sky.cs.berkeley.edu/people/); Matei Zaharia (Core faculty; https://sky.cs.berkeley.edu/people/)
- Landmark 2017: Ray: A Distributed Framework for Emerging AI Applications (https://arxiv.org/abs/1712.05889)
- Landmark 2023: Efficient Memory Management for Large Language Model Serving with PagedAttention (vLLM) (https://arxiv.org/abs/2309.06180)
- Landmark 2023: SkyPilot: An Intercloud Broker for Sky Computing (https://www.usenix.org/conference/nsdi23/presentation/yang-zongheng)
- Landmark 2024: Chatbot Arena: An Open Platform for Evaluating LLMs by Human Preference (https://arxiv.org/abs/2403.04132)
Links: https://github.com/skypilot-org
Sources: https://sky.cs.berkeley.edu/

## Black Forest Labs
https://labs.fru.dev/labs/black-forest-labs | Startup lab | founded 2024 | Freiburg, Germany | https://bfl.ai
Known for: The FLUX image models, from the researchers behind latent diffusion and Stable Diffusion.
Leaders: Robin Rombach (Co-founder; https://bfl.ai/about); Andreas Blattmann (Co-founder; https://bfl.ai/about); Patrick Esser (Co-founder; https://bfl.ai/about)
Flagship model: FLUX (https://bfl.ai/)
- Landmark 2021: High-Resolution Image Synthesis with Latent Diffusion Models (https://arxiv.org/abs/2112.10752)
- Landmark 2025: FLUX.1 Kontext: Flow Matching for In-Context Image Generation and Editing in Latent Space (https://arxiv.org/abs/2506.15742)
Links: https://github.com/black-forest-labs https://huggingface.co/black-forest-labs https://x.com/bfl_ml
Sources: https://www.wikidata.org/wiki/Q128801641 https://bfl.ai/about

## ByteDance Seed
https://labs.fru.dev/labs/bytedance-seed | Company lab of ByteDance | China | https://seed.bytedance.com
Known for: ByteDance's model team behind Doubao, the UI-TARS computer-use agents and Seedance video generation.
- Landmark 2025: UI-TARS: Pioneering Automated GUI Interaction with Native Agents (https://arxiv.org/abs/2501.12326)
- Landmark 2025: DAPO: An Open-Source LLM Reinforcement Learning System at Scale (https://arxiv.org/abs/2503.14476)
- Landmark 2025: Seed1.5-VL Technical Report (https://arxiv.org/abs/2505.07062)
- Landmark 2025: Seedance 1.0: Exploring the Boundaries of Video Generation Models (https://arxiv.org/abs/2506.09113)
Links: https://github.com/ByteDance-Seed https://huggingface.co/ByteDance-Seed
Sources: https://seed.bytedance.com/en/

## CMU Database Group
https://labs.fru.dev/labs/cmu-database-group | Academic of Carnegie Mellon University | Pittsburgh, United States | https://db.cs.cmu.edu
Known for: Database systems research and teaching in the open: self-driving databases, NoisePage and the Database of Databases.
Leaders: Andy Pavlo (Primary faculty; https://db.cs.cmu.edu/people/)
- Landmark 2017: Self-Driving Database Management Systems (https://www.cidrdb.org/cidr2017/papers/p42-pavlo-cidr17.pdf)
- Landmark 2016: What's Really New with NewSQL? (https://db.cs.cmu.edu/papers/2016/pavlo-newsql-sigmodrec2016.pdf)
- Landmark 2017: Database of Databases (https://dbdb.io)
Links: https://github.com/cmu-db https://x.com/CMUDB
Sources: https://db.cs.cmu.edu/people/

## CMU Machine Learning Department
https://labs.fru.dev/labs/cmu-ml | Academic of Carnegie Mellon University | founded 2006 | Pittsburgh, United States | https://www.ml.cmu.edu
Known for: The world's first machine learning department: Libratus poker, Transformer-XL, XLNet and Mamba.
Leaders: Zico Kolter (Department Head; https://www.ml.cmu.edu/about/index.html)
- Landmark 2017: Superhuman AI for heads-up no-limit poker: Libratus beats top professionals (https://www.science.org/doi/10.1126/science.aao1733)
- Landmark 2018: DARTS: Differentiable Architecture Search (https://arxiv.org/abs/1806.09055)
- Landmark 2019: Transformer-XL: Attentive Language Models Beyond a Fixed-Length Context (https://arxiv.org/abs/1901.02860)
- Landmark 2019: XLNet: Generalized Autoregressive Pretraining for Language Understanding (https://arxiv.org/abs/1906.08237)
- Landmark 2023: Mamba: Linear-Time Sequence Modeling with Selective State Spaces (https://arxiv.org/abs/2312.00752)
Links: https://x.com/mldcmu
Sources: https://www.ml.cmu.edu/about/index.html https://www.wikidata.org/wiki/Q59760081

## Cohere and Cohere Labs
https://labs.fru.dev/labs/cohere | Startup lab | founded 2019 | Toronto, Canada | https://cohere.com
Known for: Enterprise models (Command, Embed, Rerank) and Cohere Labs' open multilingual Aya research.
Leaders: Aidan Gomez (Chief Executive Officer and co-founder; https://cohere.com/about); Joelle Pineau (Chief AI Officer; https://techcrunch.com/2025/08/14/cohere-hires-long-time-meta-research-head-joelle-pineau-as-its-chief-ai-officer/); Phil Blunsom (Chief scientist; https://www.aibase.com/news/2722)
Flagship model: Command A (https://cohere.com/blog/command-a)
- Landmark 2024: Aya Model: An Instruction Finetuned Open-Access Multilingual Language Model (https://arxiv.org/abs/2402.07827)
- Landmark 2024: Aya Expanse: Combining Research Breakthroughs for a New Multilingual Frontier (https://arxiv.org/abs/2412.04261)
Links: https://github.com/cohere-ai https://huggingface.co/CohereLabs https://x.com/cohere
Sources: https://www.wikidata.org/wiki/Q110363143

## CWI Database Architectures
https://labs.fru.dev/labs/cwi-database-architectures | Academic of CWI | Amsterdam, Netherlands | https://www.cwi.nl/en/groups/database-architectures/
Known for: Column stores and vectorized execution: MonetDB, X100 (Vectorwise) and the research that became DuckDB.
Leaders: Peter Boncz (Group leader; https://www.cwi.nl/en/groups/database-architectures/)
- Landmark 2005: MonetDB/X100: Hyper-Pipelining Query Execution (https://www.cidrdb.org/cidr2005/papers/P19.pdf)
- Landmark 2004: MonetDB (https://www.monetdb.org/)
- Landmark 2019: DuckDB: an Embeddable Analytical Database (SIGMOD demo) (https://duckdb.org/pdf/SIGMOD2019-demo-duckdb.pdf)
Links: https://github.com/cwida
Sources: https://www.cwi.nl/en/groups/database-architectures/ https://www.wikidata.org/wiki/Q1054410

## Databricks Mosaic Research
https://labs.fru.dev/labs/databricks-mosaic-research | Company lab of Databricks | founded 2023 | San Francisco, United States | https://www.databricks.com/research/mosaic
Known for: Open models (MPT, DBRX) and training at scale, from the MosaicML team Databricks bought in 2023; Spark lineage.
Leaders: Ali Ghodsi (Chief Executive Officer and co-founder; https://en.wikipedia.org/wiki/Ali_Ghodsi); Matei Zaharia (Chief Technology Officer and co-founder; https://en.wikipedia.org/wiki/Matei_Zaharia)
- Landmark 2012: Resilient Distributed Datasets (Spark) (https://www.usenix.org/conference/nsdi12/technical-sessions/presentation/zaharia)
- Landmark 2023: Introducing MPT-7B (https://www.databricks.com/blog/mpt-7b)
- Landmark 2024: Introducing DBRX: A New State-of-the-Art Open LLM (https://www.databricks.com/blog/introducing-dbrx-new-state-art-open-llm)
- Landmark 2025: TAO: Using test-time compute to train efficient LLMs without labeled data (https://www.databricks.com/blog/tao-using-test-time-compute-train-efficient-llms-without-labeled-data)
Links: https://github.com/databricks https://huggingface.co/databricks https://x.com/databricks
Sources: https://www.wikidata.org/wiki/Q18350420 https://www.databricks.com/company/newsroom/press-releases/databricks-completes-acquisition-mosaicml

## DeepSeek
https://labs.fru.dev/labs/deepseek | Startup lab of High-Flyer | founded 2023 | Hangzhou, China | https://www.deepseek.com
Known for: Frontier-class open-weight models trained cheaply: DeepSeek-V3's efficient MoE and R1's open reasoning recipe.
Leaders: Liang Wenfeng (Founder; https://www.wikidata.org/wiki/Q131577453)
Flagship model: DeepSeek V4.1 Flash (https://models.fru.dev/models/deepseek-v4-1-flash)
- Landmark 2024: DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models (GRPO) (https://arxiv.org/abs/2402.03300)
- Landmark 2024: DeepSeek-V2: A Strong, Economical, and Efficient Mixture-of-Experts Language Model (https://arxiv.org/abs/2405.04434)
- Landmark 2024: DeepSeek-V3 Technical Report (https://arxiv.org/abs/2412.19437)
- Landmark 2025: DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning (https://arxiv.org/abs/2501.12948)
Links: https://github.com/deepseek-ai https://huggingface.co/deepseek-ai https://x.com/deepseek_ai
Sources: https://www.wikidata.org/wiki/Q131577453

## DuckLabs (DuckDB Labs)
https://labs.fru.dev/labs/ducklabs | Startup lab | Netherlands | https://ducklabs.com
Known for: The company of DuckDB's creators, the in-process analytics database born at CWI; Amazon agreed to buy it in 2026.
Leaders: Hannes Muhleisen (Co-founder and CEO; https://duckdblabs.com/about/)
- Landmark 2019: DuckDB: an Embeddable Analytical Database (SIGMOD demo) (https://duckdb.org/pdf/SIGMOD2019-demo-duckdb.pdf)
- Landmark 2024: Announcing DuckDB 1.0.0 (https://duckdb.org/2024/06/03/announcing-duckdb-100)
- Landmark 2025: DuckLake: SQL as a Lakehouse Format (https://duckdb.org/2025/05/27/ducklake)
- Funding 2026-08-26: Acquisition by Amazon (announced), the team joins Amazon Web Services (https://www.geekwire.com/2026/amazon-acquires-ducklabs-adding-the-team-behind-duckdb-amid-broader-shakeup-in-cloud-data/)
Links: https://github.com/duckdb https://x.com/duckdb
Sources: https://duckdblabs.com/about/ https://en.wikipedia.org/wiki/DuckDB

## EleutherAI
https://labs.fru.dev/labs/eleutherai | Nonprofit | founded 2020 | United States | https://www.eleuther.ai
Known for: Grassroots open research: The Pile, GPT-NeoX and Pythia, and the LM Evaluation Harness everyone uses.
Leaders: Stella Biderman (Executive Director; https://www.eleuther.ai/staff)
- Landmark 2020: The Pile: An 800GB Dataset of Diverse Text for Language Modeling (https://arxiv.org/abs/2101.00027)
- Landmark 2022: GPT-NeoX-20B: An Open-Source Autoregressive Language Model (https://arxiv.org/abs/2204.06745)
- Landmark 2023: Pythia: A Suite for Analyzing Large Language Models Across Training and Scaling (https://arxiv.org/abs/2304.01373)
- Landmark 2021: Language Model Evaluation Harness (https://github.com/EleutherAI/lm-evaluation-harness)
Links: https://github.com/EleutherAI https://huggingface.co/EleutherAI https://x.com/AiEleuther
Sources: https://www.wikidata.org/wiki/Q106289326 https://en.wikipedia.org/wiki/EleutherAI

## Epoch AI
https://labs.fru.dev/labs/epoch-ai | Nonprofit | founded 2022 | United States | https://epoch.ai
Known for: Data on AI's trajectory: training compute trends, data limits, and the FrontierMath benchmark.
Leaders: Jaime Sevilla (Co-founder and CEO; https://epoch.ai/about/team/jaime-sevilla)
- Landmark 2022: Compute Trends Across Three Eras of Machine Learning (https://arxiv.org/abs/2202.05924)
- Landmark 2022: Will we run out of data? Limits of LLM scaling based on human-generated data (https://arxiv.org/abs/2211.04325)
- Landmark 2024: FrontierMath: A Benchmark for Evaluating Advanced Mathematical Reasoning in AI (https://arxiv.org/abs/2411.04872)
Links: https://github.com/epoch-research https://x.com/EpochAIResearch
Sources: https://epoch.ai/about

## ETH AI Center
https://labs.fru.dev/labs/eth-ai-center | Academic of ETH Zurich | founded 2020 | Zurich, Switzerland | https://ai.ethz.ch
Known for: ETH Zurich's cross-department AI hub and a driver of the Swiss AI Initiative's open Apertus model.
Leaders: Alexander Ilic (Co-Founder and Executive Director; https://ai.ethz.ch/about-us.html); Andreas Krause (Chair of the steering committee; https://ai.ethz.ch/about-us.html)
- Landmark 2025: Apertus: Switzerland's open large language model (https://ethz.ch/en/news-and-events/eth-news/news/2025/09/press-release-apertus-a-fully-open-transparent-multilingual-language-model.html)
Links: https://github.com/swiss-ai https://huggingface.co/swiss-ai
Sources: https://ai.ethz.ch/about-us.html

## Google DeepMind
https://labs.fru.dev/labs/google-deepmind | Company lab of Google | founded 2010 | London, United Kingdom | https://deepmind.google
Known for: AlphaGo and AlphaFold, deep reinforcement learning, and Google's Gemini models.
Leaders: Noam Shazeer (Co-lead, Gemini; https://en.wikipedia.org/wiki/Noam_Shazeer); Demis Hassabis (Chief Executive Officer, Google DeepMind; https://en.wikipedia.org/wiki/Demis_Hassabis); Koray Kavukcuoglu (chief AI architect; https://www.semafor.com/article/06/11/2025/google-names-new-chief-ai-architect-to-advance-developments)
Flagship model: Gemini 3.8 Flash (https://models.fru.dev/models/gemini-3-8-flash)
- Landmark 2015: Human-level control through deep reinforcement learning (DQN) (https://www.nature.com/articles/nature14236)
- Landmark 2016: Mastering the game of Go with deep neural networks and tree search (AlphaGo) (https://www.nature.com/articles/nature16961)
- Landmark 2021: Highly accurate protein structure prediction with AlphaFold (https://www.nature.com/articles/s41586-021-03819-2)
- Landmark 2022: Training Compute-Optimal Large Language Models (Chinchilla) (https://arxiv.org/abs/2203.15556)
- Landmark 2023: Gemini: A Family of Highly Capable Multimodal Models (https://arxiv.org/abs/2312.11805)
Links: https://github.com/google-deepmind https://huggingface.co/google https://x.com/GoogleDeepMind
Sources: https://www.wikidata.org/wiki/Q15733006

## Google Research
https://labs.fru.dev/labs/google-research | Company lab of Google | founded 2000 | Mountain View, United States | https://research.google
Known for: The Transformer, BERT and T5, plus MapReduce and TensorFlow: much of the base of modern AI.
Leaders: Yossi Matias (Vice President, Google, and Head of Google Research; https://research.google/)
- Landmark 2004: MapReduce: Simplified Data Processing on Large Clusters (https://research.google/pubs/mapreduce-simplified-data-processing-on-large-clusters/)
- Landmark 2017: Attention Is All You Need (https://arxiv.org/abs/1706.03762)
- Landmark 2018: BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding (https://arxiv.org/abs/1810.04805)
- Landmark 2019: Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer (T5) (https://arxiv.org/abs/1910.10683)
- Landmark 2022: Chain-of-Thought Prompting Elicits Reasoning in Large Language Models (https://arxiv.org/abs/2201.11903)
Links: https://github.com/google-research https://huggingface.co/google https://x.com/GoogleResearch
Sources: https://www.wikidata.org/wiki/Q28943742 https://en.wikipedia.org/wiki/Google_Research https://research.google/

## Hugging Face
https://labs.fru.dev/labs/hugging-face | Startup lab | founded 2016 | New York, United States | https://huggingface.co
Known for: The Hub for open models and datasets and the Transformers library; also BLOOM, StarCoder, FineWeb and SmolLM.
Leaders: Clément Delangue (Chief Executive Officer and co-founder; https://en.wikipedia.org/wiki/Hugging_Face); Julien Chaumond (Chief Technology Officer and co-founder; https://en.wikipedia.org/wiki/Hugging_Face); Thomas Wolf (Chief Science Officer and co-founder; https://en.wikipedia.org/wiki/Hugging_Face)
- Landmark 2019: HuggingFace's Transformers: State-of-the-art Natural Language Processing (https://arxiv.org/abs/1910.03771)
- Landmark 2022: BLOOM: A 176B-Parameter Open-Access Multilingual Language Model (https://arxiv.org/abs/2211.05100)
- Landmark 2023: StarCoder: may the source be with you! (https://arxiv.org/abs/2305.06161)
- Landmark 2024: The FineWeb Datasets: Decanting the Web for the Finest Text Data at Scale (https://arxiv.org/abs/2406.17557)
- Landmark 2025: SmolLM2: When Smol Goes Big (https://arxiv.org/abs/2502.02737)
- Funding 2026-09-03: Acquisition by NVIDIA (agreed) $12.93B, Hugging Face stays an open platform, per NVIDIA (https://blogs.nvidia.com/blog/nvidia-to-acquire-hugging-face/)
Links: https://github.com/huggingface https://huggingface.co/huggingface https://x.com/huggingface
Sources: https://www.wikidata.org/wiki/Q108943604 https://en.wikipedia.org/wiki/Hugging_Face

## IBM Research
https://labs.fru.dev/labs/ibm-research | Company lab of IBM | founded 1945 | Yorktown Heights, United States | https://research.ibm.com
Known for: Deep Blue and Watson, relational databases and quantum computing, and today the open Granite models.
Leaders: Jay Gambetta (Director of Research and IBM Fellow; https://research.ibm.com/people/jay-gambetta)
Flagship model: IBM Granite (https://www.ibm.com/granite)
- Landmark 1970: A Relational Model of Data for Large Shared Data Banks (https://dl.acm.org/doi/10.1145/362384.362685)
- Landmark 2024: Granite Code Models: A Family of Open Foundation Models for Code Intelligence (https://arxiv.org/abs/2405.04324)
- Landmark 2024: Docling Technical Report (https://arxiv.org/abs/2408.09869)
Links: https://github.com/IBM https://huggingface.co/ibm-granite https://x.com/IBMResearch
Sources: https://www.wikidata.org/wiki/Q3146518 https://en.wikipedia.org/wiki/IBM_Research

## Isomorphic Labs
https://labs.fru.dev/labs/isomorphic-labs | Company lab of Alphabet | founded 2021 | London, United Kingdom | https://www.isomorphiclabs.com
Known for: AI-first drug discovery spun out of DeepMind; co-developed AlphaFold 3.
Leaders: Demis Hassabis (Founder and CEO; https://www.isomorphiclabs.com/)
- Landmark 2024: Accurate structure prediction of biomolecular interactions with AlphaFold 3 (https://www.nature.com/articles/s41586-024-07487-w)
Links: https://x.com/IsomorphicLabs
Sources: https://www.wikidata.org/wiki/Q109536010 https://en.wikipedia.org/wiki/Isomorphic_Labs

## Kyutai
https://labs.fru.dev/labs/kyutai | Nonprofit | Paris, France | https://kyutai.org
Known for: Open-science lab funded by Xavier Niel and partners: Moshi real-time voice and Hibiki live translation.
- Landmark 2024: Moshi: a speech-text foundation model for real-time dialogue (https://arxiv.org/abs/2410.00037)
- Landmark 2025: High-Fidelity Simultaneous Speech-To-Speech Translation (Hibiki) (https://arxiv.org/abs/2502.03382)
Links: https://github.com/kyutai-labs https://huggingface.co/kyutai https://x.com/kyutai_labs
Sources: https://kyutai.org/ https://en.wikipedia.org/wiki/Xavier_Niel

## LAION
https://labs.fru.dev/labs/laion | Nonprofit | Germany | https://laion.ai
Known for: Open image-text datasets (LAION-400M, LAION-5B) that trained Stable Diffusion and OpenCLIP.
- Landmark 2021: LAION-400M: Open Dataset of CLIP-Filtered 400 Million Image-Text Pairs (https://arxiv.org/abs/2111.02114)
- Landmark 2022: LAION-5B: An open large-scale dataset for training next generation image-text models (https://arxiv.org/abs/2210.08402)
- Landmark 2022: Reproducible scaling laws for contrastive language-image learning (https://arxiv.org/abs/2212.07143)
Links: https://github.com/LAION-AI https://huggingface.co/laion https://x.com/laion_ai
Sources: https://laion.ai/about/ https://www.wikidata.org/wiki/Q114586027

## Liquid AI
https://labs.fru.dev/labs/liquid-ai | Startup lab | founded 2023 | Cambridge, Massachusetts, United States | https://www.liquid.ai
Known for: Liquid foundation models (LFM) for devices, built on liquid neural network research from MIT CSAIL.
Leaders: Ramin Hasani (CEO and co-founder; https://www.liquid.ai/company)
- Landmark 2020: Liquid Time-constant Networks (https://arxiv.org/abs/2006.04439)
Links: https://github.com/Liquid4All https://huggingface.co/LiquidAI https://x.com/liquidai
Sources: https://www.liquid.ai/company https://en.wikipedia.org/wiki/Liquid_AI

## Max Planck Institute for Intelligent Systems
https://labs.fru.dev/labs/mpi-is | Academic of Max Planck Society | founded 2011 | Stuttgart, Germany | https://is.mpg.de
Known for: Causal inference and kernel methods (Scholkopf), and the SMPL body model used across graphics and vision.
- Landmark 2015: SMPL: A Skinned Multi-Person Linear Model (https://smpl.is.tue.mpg.de/)
- Landmark 2021: Towards Causal Representation Learning (https://arxiv.org/abs/2102.11107)
Links: https://x.com/MPI_IS
Sources: https://www.wikidata.org/wiki/Q1287942 https://en.wikipedia.org/wiki/Max_Planck_Institute_for_Intelligent_Systems

## Meta FAIR
https://labs.fru.dev/labs/meta-fair | Company lab of Meta | founded 2013 | New York, United States | https://ai.meta.com/research/
Known for: PyTorch, the open Llama models, Segment Anything and self-supervised vision (DINO).
- Landmark 2017: Billion-scale similarity search with GPUs (Faiss) (https://arxiv.org/abs/1702.08734)
- Landmark 2019: PyTorch: An Imperative Style, High-Performance Deep Learning Library (https://arxiv.org/abs/1912.01703)
- Landmark 2023: LLaMA: Open and Efficient Foundation Language Models (https://arxiv.org/abs/2302.13971)
- Landmark 2023: Segment Anything (https://arxiv.org/abs/2304.02643)
- Landmark 2023: DINOv2: Learning Robust Visual Features without Supervision (https://arxiv.org/abs/2304.07193)
Links: https://github.com/facebookresearch https://huggingface.co/facebook https://x.com/AIatMeta
Sources: https://www.wikidata.org/wiki/Q112114913 https://en.wikipedia.org/wiki/Meta_AI https://ai.meta.com/research/

## Meta Superintelligence Labs
https://labs.fru.dev/labs/meta-superintelligence-labs | Company lab of Meta | founded 2025 | Menlo Park, United States | https://ai.meta.com
Known for: Meta's frontier model effort, formed in 2025 under Alexandr Wang; it builds the Muse models and now houses FAIR.
Leaders: Alexandr Wang (Chief AI Officer, leads Meta Superintelligence Labs; https://www.cnbc.com/2025/06/30/mark-zuckerberg-creating-meta-superintelligence-labs-read-the-memo.html); Shengjia Zhao (Chief Scientist of Meta Superintelligence Labs; https://techcrunch.com/2025/07/25/meta-names-shengjia-zhao-as-chief-scientist-of-ai-superintelligence-unit/); Nat Friedman (Head of AI products and applied research, Meta Superintelligence Labs; https://www.cnbc.com/2025/06/30/mark-zuckerberg-creating-meta-superintelligence-labs-read-the-memo.html)
Flagship model: Muse Spark 1.3 (https://models.fru.dev/models/muse-spark-1-3)
- Landmark 2024: The Llama 3 Herd of Models (https://arxiv.org/abs/2407.21783)
Links: https://github.com/meta-llama https://huggingface.co/meta-llama https://x.com/AIatMeta
Sources: https://www.wikidata.org/wiki/Q135397403 https://en.wikipedia.org/wiki/Meta_Superintelligence_Labs

## METR
https://labs.fru.dev/labs/metr | Nonprofit | founded 2022 | Berkeley, California, United States | https://metr.org
Known for: Evaluating dangerous capabilities: the time-horizon measure of how long a task AI agents can finish.
Leaders: Beth Barnes (Founder and CEO; https://metr.org/about)
- Landmark 2024: RE-Bench: Evaluating frontier AI R&D capabilities of language model agents against human experts (https://arxiv.org/abs/2411.15114)
- Landmark 2025: Measuring AI Ability to Complete Long Tasks (https://arxiv.org/abs/2503.14499)
- Landmark 2025: Measuring the Impact of Early-2025 AI on Experienced Open-Source Developer Productivity (https://arxiv.org/abs/2507.09089)
Links: https://github.com/METR https://x.com/METR_Evals
Sources: https://www.wikidata.org/wiki/Q135185153 https://en.wikipedia.org/wiki/METR https://metr.org/about

## Microsoft AI
https://labs.fru.dev/labs/microsoft-ai | Company lab of Microsoft | founded 2024 | Redmond, United States | https://microsoft.ai
Known for: Copilot for consumers and Microsoft's own MAI models, led by Mustafa Suleyman since 2024.
Leaders: Mustafa Suleyman (Executive Vice President and CEO, Microsoft AI; https://blogs.microsoft.com/blog/2024/03/19/mustafa-suleyman-deepmind-and-inflection-co-founder-joins-microsoft-to-lead-copilot/)
- Landmark 2024: Mustafa Suleyman joins Microsoft to lead Copilot (https://blogs.microsoft.com/blog/2024/03/19/mustafa-suleyman-deepmind-and-inflection-co-founder-joins-microsoft-to-lead-copilot/)
Links: https://github.com/microsoft https://huggingface.co/microsoft
Sources: https://www.wikidata.org/wiki/Q125891217 https://en.wikipedia.org/wiki/Microsoft_AI https://blogs.microsoft.com/blog/2024/03/19/mustafa-suleyman-deepmind-and-inflection-co-founder-joins-microsoft-to-lead-copilot/

## Microsoft Research
https://labs.fru.dev/labs/microsoft-research | Company lab of Microsoft | founded 1991 | Redmond, United States | https://www.microsoft.com/en-us/research/
Known for: ResNet, LoRA, DeepSpeed and the small Phi models, from a lab spanning systems, theory and AI.
Leaders: Peter Lee (President, Microsoft Research; https://www.microsoft.com/en-us/research/about-microsoft-research/)
Flagship model: Phi-4 (https://arxiv.org/abs/2412.08905)
- Landmark 2015: Deep Residual Learning for Image Recognition (ResNet) (https://arxiv.org/abs/1512.03385)
- Landmark 2019: ZeRO: Memory Optimizations Toward Training Trillion Parameter Models (https://arxiv.org/abs/1910.02054)
- Landmark 2021: LoRA: Low-Rank Adaptation of Large Language Models (https://arxiv.org/abs/2106.09685)
- Landmark 2023: Textbooks Are All You Need (phi-1) (https://arxiv.org/abs/2306.11644)
- Landmark 2024: From Local to Global: A Graph RAG Approach to Query-Focused Summarization (https://arxiv.org/abs/2404.16130)
Links: https://github.com/microsoft https://huggingface.co/microsoft https://x.com/MSFTResearch
Sources: https://www.wikidata.org/wiki/Q1144725 https://www.microsoft.com/en-us/research/about-microsoft-research/

## Mila
https://labs.fru.dev/labs/mila | Academic | founded 1993 | Montreal, Canada | https://mila.quebec
Known for: Yoshua Bengio's institute: neural language models, attention for translation, GANs and Theano.
Leaders: Hugo Larochelle (Scientific Director; https://mila.quebec/en/directory/hugo-larochelle)
- Landmark 2003: A Neural Probabilistic Language Model (https://www.jmlr.org/papers/v3/bengio03a.html)
- Landmark 2014: Generative Adversarial Networks (https://arxiv.org/abs/1406.2661)
- Landmark 2014: Neural Machine Translation by Jointly Learning to Align and Translate (https://arxiv.org/abs/1409.0473)
- Landmark 2016: Theano: A Python framework for fast computation of mathematical expressions (https://arxiv.org/abs/1605.02688)
Links: https://github.com/mila-iqia https://x.com/Mila_Quebec
Sources: https://mila.quebec/en/mila

## MiniMax
https://labs.fru.dev/labs/minimax | Startup lab | founded 2021 | Shanghai, China | https://www.minimax.io
Known for: Lightning attention for million-token context in open models, plus Hailuo video and speech models.
Leaders: Yan Junjie (Founder; https://www.wikidata.org/wiki/Q130263208)
- Landmark 2025: MiniMax-01: Scaling Foundation Models with Lightning Attention (https://arxiv.org/abs/2501.08313)
- Landmark 2025: MiniMax-M1: Scaling Test-Time Compute Efficiently with Lightning Attention (https://arxiv.org/abs/2506.13585)
- Funding 2026-01-09: IPO, listed on the Hong Kong Stock Exchange (https://www.scmp.com/business/banking-finance/article/3339251/chinese-ai-start-minimax-shines-hong-kong-ipo-debut)
Links: https://github.com/MiniMax-AI https://huggingface.co/MiniMaxAI https://x.com/MiniMax__AI
Sources: https://www.wikidata.org/wiki/Q130263208 https://en.wikipedia.org/wiki/MiniMax_Group

## Mistral AI
https://labs.fru.dev/labs/mistral-ai | Startup lab | founded 2023 | Paris, France | https://mistral.ai
Known for: Europe's leading model maker: efficient open-weight models, Mixtral's mixture of experts, and Le Chat.
Leaders: Arthur Mensch (Chief Executive Officer and co-founder; https://en.wikipedia.org/wiki/Mistral_AI); Timothée Lacroix (CTO; https://en.wikipedia.org/wiki/Mistral_AI)
Flagship model: Mistral Medium 3 (https://mistral.ai/news/mistral-medium-3)
- Landmark 2023: Mistral 7B (https://arxiv.org/abs/2310.06825)
- Landmark 2024: Mixtral of Experts (https://arxiv.org/abs/2401.04088)
- Landmark 2025: Magistral (https://arxiv.org/abs/2506.10910)
- Funding 2026-09-08: Series D $3.6B, at a $24.4B valuation (https://techcrunch.com/2026/09/08/mistral-raises-e3b-as-sovereign-ai-becomes-big-business/)
Links: https://github.com/mistralai https://huggingface.co/mistralai https://x.com/MistralAI
Sources: https://www.wikidata.org/wiki/Q119718658

## MIT CSAIL
https://labs.fru.dev/labs/mit-csail | Academic of MIT | founded 2003 | Cambridge, Massachusetts, United States | https://www.csail.mit.edu
Known for: MIT's largest lab: robotics, the Julia language, adversarial robustness and the lottery ticket hypothesis.
Leaders: Daniela Rus (Director; https://www.csail.mit.edu/about/leadership)
- Landmark 2012: Julia: A Fast Dynamic Language for Technical Computing (https://arxiv.org/abs/1209.5145)
- Landmark 2017: Towards Deep Learning Models Resistant to Adversarial Attacks (https://arxiv.org/abs/1706.06083)
- Landmark 2018: The Lottery Ticket Hypothesis: Finding Sparse, Trainable Neural Networks (https://arxiv.org/abs/1803.03635)
- Landmark 2020: Liquid Time-constant Networks (https://arxiv.org/abs/2006.04439)
Links: https://x.com/MIT_CSAIL
Sources: https://www.wikidata.org/wiki/Q1354917 https://www.csail.mit.edu/about/leadership

## Moonshot AI
https://labs.fru.dev/labs/moonshot-ai | Startup lab | founded 2023 | Beijing, China | https://www.moonshot.ai
Known for: The Kimi models: long context early, then the open-weight Kimi K2 trained with the Muon optimizer.
Leaders: Yang Zhilin (Chief Executive Officer; https://en.wikipedia.org/wiki/Moonshot_AI)
Flagship model: Kimi K3 (https://models.fru.dev/models/kimi-k3)
- Landmark 2025: Kimi k1.5: Scaling Reinforcement Learning with LLMs (https://arxiv.org/abs/2501.12599)
- Landmark 2025: Muon is Scalable for LLM Training (https://arxiv.org/abs/2502.16982)
- Landmark 2025: Kimi K2: Open Agentic Intelligence (https://arxiv.org/abs/2507.20534)
- Funding 2026-07-29: Growth $3.5B, at a $35B valuation (https://finance.yahoo.com/technology/ai/articles/moonshot-ai-3-5-billion-072537342.html)
Links: https://github.com/MoonshotAI https://huggingface.co/moonshotai https://x.com/Kimi_Moonshot
Sources: https://www.wikidata.org/wiki/Q130270266

## Nous Research
https://labs.fru.dev/labs/nous-research | Startup lab | United States | https://nousresearch.com
Known for: The open Hermes fine-tunes, YaRN context extension, and decentralized training over the internet.
- Landmark 2023: YaRN: Efficient Context Window Extension of Large Language Models (https://arxiv.org/abs/2309.00071)
- Landmark 2024: Hermes 3 Technical Report (https://arxiv.org/abs/2408.11857)
- Landmark 2024: DisTrO: distributed training over the internet (https://github.com/NousResearch/DisTrO)
Links: https://github.com/NousResearch https://huggingface.co/NousResearch https://x.com/NousResearch
Sources: https://nousresearch.com/ https://github.com/NousResearch

## NVIDIA Research
https://labs.fru.dev/labs/nvidia-research | Company lab of NVIDIA | founded 2006 | Santa Clara, United States | https://research.nvidia.com
Known for: StyleGAN, Megatron-LM for training giant models, Instant NeRF, and the open Nemotron models.
Leaders: Bill Dally (Chief Scientist and Senior Vice President of Research; https://en.wikipedia.org/wiki/Bill_Dally)
Flagship model: Nemotron-4 340B (https://arxiv.org/abs/2406.11704)
- Landmark 2018: A Style-Based Generator Architecture for Generative Adversarial Networks (StyleGAN) (https://arxiv.org/abs/1812.04948)
- Landmark 2019: Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism (https://arxiv.org/abs/1909.08053)
- Landmark 2022: Instant Neural Graphics Primitives with a Multiresolution Hash Encoding (https://arxiv.org/abs/2201.05989)
- Landmark 2025: Cosmos World Foundation Model Platform for Physical AI (https://arxiv.org/abs/2501.03575)
Links: https://github.com/NVlabs https://huggingface.co/nvidia
Sources: https://research.nvidia.com/about

## OpenAI
https://labs.fru.dev/labs/openai | Startup lab | founded 2015 | San Francisco, United States | https://openai.com
Known for: The GPT models and ChatGPT, and the scaling results that set off the large language model race.
Leaders: Sam Altman (Chief Executive Officer; https://en.wikipedia.org/wiki/OpenAI); Vijaye Raji (CTO of Applications; https://techcrunch.com/2025/09/02/openai-acquires-product-testing-startup-statsig-and-shakes-up-its-leadership-team/); Jakub Pachocki (Chief Scientist; https://www.theverge.com/2024/5/14/24156920/openai-chief-scientist-ilya-sutskever-leaves); Sachin Katti (Leads compute infrastructure for AGI; https://www.crn.com/news/components-peripherals/2025/intel-ai-leader-sachin-katti-decamps-to-openai)
Flagship model: GPT-6 Astra (https://models.fru.dev/models/gpt-6-astra)
- Landmark 2020: Scaling Laws for Neural Language Models (https://arxiv.org/abs/2001.08361)
- Landmark 2020: Language Models are Few-Shot Learners (GPT-3) (https://arxiv.org/abs/2005.14165)
- Landmark 2021: Learning Transferable Visual Models From Natural Language Supervision (CLIP) (https://arxiv.org/abs/2103.00020)
- Landmark 2022: Training language models to follow instructions with human feedback (InstructGPT) (https://arxiv.org/abs/2203.02155)
- Landmark 2022: Introducing ChatGPT (https://openai.com/index/chatgpt/)
- Funding 2026-03-31: Private round $122B, at an $852B post-money valuation (https://openai.com/index/accelerating-the-next-phase-ai/)
- Compute 2025-01-21: Stargate (SoftBank, Oracle, MGX), The Stargate Project: up to $500B of US AI infrastructure over four years (https://openai.com/index/announcing-the-stargate-project/)
- Compute 2025-09-22: NVIDIA, At least 10 GW of NVIDIA systems (https://openai.com/index/openai-nvidia-systems-partnership/)
- Compute 2025-10-06: AMD, 6 GW of AMD Instinct GPUs (https://openai.com/index/openai-amd-strategic-partnership/)
- Compute 2025-10-13: Broadcom, 10 GW of custom AI accelerators (https://openai.com/index/openai-and-broadcom-announce-strategic-collaboration/)
- Compute 2025-11-03: Amazon Web Services, Multi-year AWS compute partnership (https://openai.com/index/aws-and-openai-partnership/)
Links: https://github.com/openai https://huggingface.co/openai https://x.com/OpenAI
Sources: https://www.wikidata.org/wiki/Q21708200

## Physical Intelligence
https://labs.fru.dev/labs/physical-intelligence | Startup lab | founded 2024 | San Francisco, United States | https://www.physicalintelligence.company
Known for: General-purpose robot foundation models: the pi-zero vision-language-action family, open-sourced as openpi.
- Landmark 2024: pi0: A Vision-Language-Action Flow Model for General Robot Control (https://arxiv.org/abs/2410.24164)
- Landmark 2025: FAST: Efficient Action Tokenization for Vision-Language-Action Models (https://arxiv.org/abs/2501.09747)
- Landmark 2025: pi0.5: a Vision-Language-Action Model with Open-World Generalization (https://arxiv.org/abs/2504.16054)
Links: https://github.com/Physical-Intelligence https://huggingface.co/physical-intelligence
Sources: https://www.wikidata.org/wiki/Q138658641 https://en.wikipedia.org/wiki/Physical_Intelligence_Inc.

## Reka
https://labs.fru.dev/labs/reka | Startup lab | founded 2022 | Sunnyvale, United States | https://reka.ai
Known for: Natively multimodal models (Core, Flash, Edge) from former DeepMind and Meta researchers.
Leaders: Dani Yogatama (Co-founder; https://reka.ai/)
- Landmark 2024: Reka Core, Flash, and Edge: A Series of Powerful Multimodal Language Models (https://arxiv.org/abs/2404.12387)
Links: https://github.com/reka-ai https://huggingface.co/RekaAI https://x.com/RekaAILabs
Sources: https://www.wikidata.org/wiki/Q125878748 https://reka.ai/

## Safe Superintelligence (SSI)
https://labs.fru.dev/labs/safe-superintelligence | Startup lab | founded 2024 | Palo Alto, United States | https://ssi.inc
Known for: Ilya Sutskever's lab with one goal and one product, a safe superintelligence; it has published nothing yet.
Leaders: Ilya Sutskever (Chief Executive Officer and co-founder; https://en.wikipedia.org/wiki/Safe_Superintelligence_Inc.)
- Landmark 2024: Safe Superintelligence Inc. founding statement (https://ssi.inc/)
- Funding 2026-07-27: Strategic $5B (https://www.calcalistech.com/ctechnews/article/qy9eg5jw4)
Sources: https://www.wikidata.org/wiki/Q130302670 https://ssi.inc/

## Sakana AI
https://labs.fru.dev/labs/sakana-ai | Startup lab | founded 2023 | Tokyo, Japan | https://sakana.ai
Known for: Nature-inspired methods: evolutionary model merging and The AI Scientist, from a Transformer co-author.
Leaders: David Ha (Co-founder; https://www.wikidata.org/wiki/Q126691075); Llion Jones (Co-founder; https://www.wikidata.org/wiki/Q126691075)
- Landmark 2024: Evolutionary Optimization of Model Merging Recipes (https://arxiv.org/abs/2403.13187)
- Landmark 2024: The AI Scientist: Towards Fully Automated Open-Ended Scientific Discovery (https://arxiv.org/abs/2408.06292)
- Landmark 2025: Darwin Godel Machine: Open-Ended Evolution of Self-Improving Agents (https://arxiv.org/abs/2505.22954)
Links: https://github.com/SakanaAI https://huggingface.co/SakanaAI https://x.com/SakanaAILabs
Sources: https://www.wikidata.org/wiki/Q126691075

## Salesforce AI Research
https://labs.fru.dev/labs/salesforce-ai-research | Company lab of Salesforce | founded 2016 | United States | https://www.salesforce.com/research/
Known for: BLIP vision-language models, CodeT5, and xLAM action models for agents.
- Landmark 2019: CTRL: A Conditional Transformer Language Model for Controllable Generation (https://arxiv.org/abs/1909.05858)
- Landmark 2021: CodeT5: Identifier-aware Unified Pre-trained Encoder-Decoder Models for Code Understanding and Generation (https://arxiv.org/abs/2109.00859)
- Landmark 2022: BLIP: Bootstrapping Language-Image Pre-training for Unified Vision-Language Understanding and Generation (https://arxiv.org/abs/2201.12086)
- Landmark 2024: Unified Training of Universal Time Series Forecasting Transformers (Moirai) (https://arxiv.org/abs/2402.02592)
- Landmark 2024: xLAM: A Family of Large Action Models to Empower AI Agent Systems (https://arxiv.org/abs/2409.03215)
Links: https://github.com/salesforce https://huggingface.co/Salesforce
Sources: https://www.salesforce.com/research/ https://en.wikipedia.org/wiki/Richard_Socher

## Shanghai AI Laboratory
https://labs.fru.dev/labs/shanghai-ai-lab | Government | Shanghai, China | https://www.shlab.org.cn
Known for: The InternLM and InternVL open models and the OpenCompass evaluation platform.
- Landmark 2023: InternVL: Scaling up Vision Foundation Models and Aligning for Generic Visual-Linguistic Tasks (https://arxiv.org/abs/2312.14238)
- Landmark 2024: InternLM2 Technical Report (https://arxiv.org/abs/2403.17297)
- Landmark 2023: OpenCompass: a platform for evaluating large models (https://github.com/open-compass/opencompass)
Links: https://github.com/InternLM https://huggingface.co/internlm
Sources: https://www.wikidata.org/wiki/Q124251933

## Snowflake AI Research
https://labs.fru.dev/labs/snowflake-ai-research | Company lab of Snowflake | Menlo Park, United States | https://www.snowflake.com/en/engineering-blog/
Known for: Arctic: open embedding models, the Arctic LLM, SwiftKV inference and text-to-SQL research.
Leaders: Sridhar Ramaswamy (Chief Executive Officer; https://en.wikipedia.org/wiki/Sridhar_Ramaswamy); Thierry Cruanes (CTO; https://en.wikipedia.org/wiki/Snowflake_Inc.)
- Landmark 2024: Snowflake Arctic: the best LLM for enterprise AI (https://www.snowflake.com/en/blog/arctic-open-efficient-foundation-language-models-snowflake/)
- Landmark 2024: Arctic-Embed: Scalable, Efficient, and Accurate Text Embedding Models (https://arxiv.org/abs/2405.05374)
- Landmark 2024: SwiftKV: Fast Prefill-Optimized Inference with Knowledge-Preserving Model Transformation (https://arxiv.org/abs/2410.03960)
Links: https://github.com/Snowflake-Labs https://huggingface.co/Snowflake https://x.com/Snowflake
Sources: https://www.wikidata.org/wiki/Q22078063 https://en.wikipedia.org/wiki/Snowflake_Inc. https://www.snowflake.com/en/engineering-blog/

## Stability AI
https://labs.fru.dev/labs/stability-ai | Startup lab | founded 2019 | London, United Kingdom | https://stability.ai
Known for: Stable Diffusion, the open image model release that opened up generative imaging in 2022.
Leaders: Prem Akkaraju (CEO; https://stability.ai/company)
- Landmark 2022: Stable Diffusion public release (https://stability.ai/news/stable-diffusion-public-release)
- Landmark 2023: SDXL: Improving Latent Diffusion Models for High-Resolution Image Synthesis (https://arxiv.org/abs/2307.01952)
- Landmark 2023: Stable Video Diffusion: Scaling Latent Video Diffusion Models to Large Datasets (https://arxiv.org/abs/2311.15127)
- Landmark 2024: Scaling Rectified Flow Transformers for High-Resolution Image Synthesis (SD3) (https://arxiv.org/abs/2403.03206)
- Funding 2026-08-25: Series B $76M (https://stability.ai/news-updates/stability-ai-latest-funding-backed-by-entertainment-industry-biggest-names)
Links: https://github.com/Stability-AI https://huggingface.co/stabilityai https://x.com/StabilityAI
Sources: https://www.wikidata.org/wiki/Q123592234 https://stability.ai/company

## Stanford HAI and CRFM
https://labs.fru.dev/labs/stanford-hai | Academic of Stanford University | founded 2019 | Stanford, California, United States | https://hai.stanford.edu
Known for: The AI Index, the term foundation models, the HELM benchmark and Alpaca, from Stanford's AI institute.
Leaders: Fei-Fei Li (Co-Director; https://www.wikidata.org/wiki/Q62171607); James Landay (Co-Director; https://hai.stanford.edu/people/james-landay)
- Landmark 2021: On the Opportunities and Risks of Foundation Models (https://arxiv.org/abs/2108.07258)
- Landmark 2022: Holistic Evaluation of Language Models (HELM) (https://arxiv.org/abs/2211.09110)
- Landmark 2023: Alpaca: A Strong, Replicable Instruction-Following Model (https://crfm.stanford.edu/2023/03/13/alpaca.html)
- Landmark 2017: The AI Index Report (https://hai.stanford.edu/ai-index)
Links: https://github.com/stanford-crfm https://huggingface.co/stanford-crfm https://x.com/StanfordHAI
Sources: https://www.wikidata.org/wiki/Q62171607

## Tencent Hy (Hunyuan)
https://labs.fru.dev/labs/tencent-hunyuan | Company lab of Tencent | founded 2023 | Shenzhen, China | https://hunyuan.tencent.com
Known for: Tencent's Hunyuan models, with open releases in MoE language, video (HunyuanVideo) and 3D generation.
- Landmark 2024: Hunyuan-Large: An Open-Source MoE Model with 52 Billion Activated Parameters by Tencent (https://arxiv.org/abs/2411.02265)
- Landmark 2024: HunyuanVideo: A Systematic Framework For Large Video Generative Models (https://arxiv.org/abs/2412.03603)
- Landmark 2025: Hunyuan3D 2.0: Scaling Diffusion Models for High Resolution Textured 3D Assets Generation (https://arxiv.org/abs/2501.12202)
Links: https://github.com/Tencent-Hunyuan https://huggingface.co/tencent
Sources: https://www.wikidata.org/wiki/Q860580 https://en.wikipedia.org/wiki/Tencent_Hy https://hunyuan.tencent.com/

## Thinking Machines Lab
https://labs.fru.dev/labs/thinking-machines-lab | Startup lab | founded 2025 | San Francisco, United States | https://thinkingmachines.ai
Known for: Mira Murati's lab: Tinker, a fine-tuning API, and open research notes on reproducible inference.
Leaders: Mira Murati (Chief Executive Officer and founder; https://en.wikipedia.org/wiki/Thinking_Machines_Lab); Soumith Chintala (CTO; https://techcrunch.com/2026/01/14/mira-muratis-startup-thinking-machines-lab-is-losing-two-of-its-co-founders-to-openai/); John Schulman (Chief Scientist; https://en.wikipedia.org/wiki/Thinking_Machines_Lab)
Flagship model: Inkling (https://models.fru.dev/models/inkling)
- Landmark 2025: Defeating Nondeterminism in LLM Inference (https://thinkingmachines.ai/blog/defeating-nondeterminism-in-llm-inference/)
- Landmark 2025: LoRA Without Regret (https://thinkingmachines.ai/blog/lora/)
- Landmark 2025: Tinker (https://thinkingmachines.ai/tinker/)
Links: https://github.com/thinking-machines-lab https://x.com/thinkymachines
Sources: https://www.wikidata.org/wiki/Q132532850 https://en.wikipedia.org/wiki/Thinking_Machines_Lab

## Tsinghua KEG
https://labs.fru.dev/labs/tsinghua-keg | Academic of Tsinghua University | Beijing, China | https://keg.cs.tsinghua.edu.cn
Known for: Tsinghua's knowledge engineering group: AMiner, the GLM and ChatGLM models (with Zhipu), CogVLM and AgentBench.
- Landmark 2008: ArnetMiner: Extraction and Mining of Academic Social Networks (https://dl.acm.org/doi/10.1145/1401890.1402008)
- Landmark 2021: GLM: General Language Model Pretraining with Autoregressive Blank Infilling (https://arxiv.org/abs/2103.10360)
- Landmark 2023: AgentBench: Evaluating LLMs as Agents (https://arxiv.org/abs/2308.03688)
- Landmark 2023: CogVLM: Visual Expert for Pretrained Language Models (https://arxiv.org/abs/2311.03079)
Links: https://github.com/THUDM https://huggingface.co/THUDM
Sources: https://keg.cs.tsinghua.edu.cn/ https://en.wikipedia.org/wiki/Z.ai

## TUM Database Systems
https://labs.fru.dev/labs/tum-database-systems | Academic of Technical University of Munich | Munich, Germany | https://db.in.tum.de
Known for: Compiled query execution in main-memory and disk-based systems (HyPer, Umbra), from Thomas Neumann's group.
Leaders: Thomas Neumann (Professor; https://db.in.tum.de/people/)
- Landmark 2011: Efficiently Compiling Efficient Query Plans for Modern Hardware (https://www.vldb.org/pvldb/vol4/p539-neumann.pdf)
- Landmark 2020: Umbra: A Disk-Based System with In-Memory Performance (https://www.cidrdb.org/cidr2020/papers/p29-neumann-cidr20.pdf)
Sources: https://db.in.tum.de/people/

## UK AI Security Institute
https://labs.fru.dev/labs/uk-aisi | Government of Department for Science, Innovation and Technology | founded 2023 | United Kingdom | https://www.aisi.gov.uk
Known for: The first state AI safety institute (renamed Security in 2025): pre-deployment testing and the Inspect eval framework.
Leaders: Henry de Zoete (Director; https://www.aisi.gov.uk/about); Ian Hogarth (Chair; https://www.aisi.gov.uk/about)
- Landmark 2024: Inspect: an open-source framework for large language model evaluations (https://inspect.aisi.org.uk/)
Links: https://github.com/UKGovernmentBEIS https://x.com/AISecurityInst
Sources: https://www.wikidata.org/wiki/Q126284597 https://en.wikipedia.org/wiki/AI_Security_Institute https://www.aisi.gov.uk/about

## US Center for AI Standards and Innovation (CAISI)
https://labs.fru.dev/labs/us-caisi | Government of NIST | founded 2023 | United States | https://www.nist.gov/caisi
Known for: NIST's AI testing center, founded as the US AI Safety Institute and renamed CAISI in 2025; evaluates frontier and foreign models.
- Landmark 2025: Center for AI Standards and Innovation (https://www.nist.gov/caisi)
Sources: https://www.wikidata.org/wiki/Q139559517 https://en.wikipedia.org/wiki/Center_for_AI_Standards_and_Innovation https://www.nist.gov/caisi

## Vector Institute
https://labs.fru.dev/labs/vector-institute | Nonprofit | founded 2017 | Toronto, Canada | https://vectorinstitute.ai
Known for: Toronto's independent AI institute, home to deep learning researchers and work such as neural ODEs.
Leaders: Glenda Crisp (President and CEO; https://vectorinstitute.ai/team/)
- Landmark 2018: Neural Ordinary Differential Equations (https://arxiv.org/abs/1806.07366)
Links: https://github.com/VectorInstitute https://x.com/VectorInst
Sources: https://www.wikidata.org/wiki/Q47462249 https://en.wikipedia.org/wiki/Vector_Institute_(Canada)

## World Labs
https://labs.fru.dev/labs/world-labs | Startup lab | San Carlos, United States | https://www.worldlabs.ai
Known for: Fei-Fei Li's spatial intelligence company: models that generate explorable 3D worlds (Marble).
Leaders: Fei-Fei Li (Co-founder; https://www.worldlabs.ai/about)
- Landmark 2024: Generating worlds (https://www.worldlabs.ai/blog)
Links: https://x.com/theworldlabs
Sources: https://www.wikidata.org/wiki/Q127527740 https://www.worldlabs.ai/about

## xAI (SpaceXAI)
https://labs.fru.dev/labs/xai | Company lab of SpaceX | founded 2023 | San Francisco Bay Area, United States | https://x.ai
Known for: The Grok models and the Colossus supercomputer in Memphis; now a division of SpaceX.
Leaders: Michael Nicolls (President, SpaceXAI; https://en.wikipedia.org/wiki/XAI_(company))
Flagship model: Grok 4.7 (https://models.fru.dev/models/grok-4-7)
- Landmark 2024: Open release of Grok-1 (https://github.com/xai-org/grok-1)
Links: https://github.com/xai-org https://huggingface.co/xai-org https://x.com/xai
Sources: https://www.wikidata.org/wiki/Q120599684 https://en.wikipedia.org/wiki/XAI_(company)

## Z.ai (Zhipu AI)
https://labs.fru.dev/labs/zhipu-ai | Startup lab | founded 2019 | Beijing, China | https://z.ai
Known for: The GLM models, spun out of Tsinghua University's KEG lab; GLM-4.5 and later are open weights.
Flagship model: GLM-5.3 (https://models.fru.dev/models/glm-5-3)
- Landmark 2021: GLM: General Language Model Pretraining with Autoregressive Blank Infilling (https://arxiv.org/abs/2103.10360)
- Landmark 2022: GLM-130B: An Open Bilingual Pre-trained Model (https://arxiv.org/abs/2210.02414)
- Landmark 2024: ChatGLM: A Family of Large Language Models from GLM-130B to GLM-4 All Tools (https://arxiv.org/abs/2406.12793)
- Landmark 2025: GLM-4.5: Agentic, Reasoning, and Coding (ARC) Foundation Models (https://arxiv.org/abs/2508.06471)
- Funding 2026-01: IPO, listed on the Hong Kong Stock Exchange (https://www.barrons.com/articles/zhipu-ai-hong-kong-ipo-china-minimax-9afc9f02)
Links: https://github.com/zai-org https://huggingface.co/zai-org https://x.com/Zai_org
Sources: https://www.wikidata.org/wiki/Q129572031 https://en.wikipedia.org/wiki/Z.ai