Changes

Builds low-latency voice AI models for conversational agents; the $100M total includes a $30M extension of its seed round.

Scale AI, Inc. is an American artificial intelligence infrastructure and software company based in San Francisco, California.

Develops AI voice generation and speech-to-text models for creators and enterprises.

Frontier AI lab building generative models of biology, known for Latent-X for de novo protein binder design.

And 164 more labs: Generate:Biomedicines, Profluent, Chai Discovery and othersShow all

• Generate:Biomedicines: Therapeutics company pioneering generative biology, known for the Chroma generative model for protein design.
• Profluent: AI protein design company that created OpenCRISPR-1, described as the first AI-designed gene editor.
• Chai Discovery: Company building AI models of biochemical structure and interaction for molecular design, starting with Chai-1.
• Recursion: AI drug discovery company that builds foundation models of cell biology from large-scale microscopy data.
• EvolutionaryScale: Frontier AI lab for the life sciences, maker of the ESM3 generative protein language model and ESM Cambrian.
• Waabi: Self-driving company founded by Raquel Urtasun, using generative AI simulation such as UniSim for autonomous trucks.
• Covariant: American AI and robotics company, formerly Embodied Intelligence, headquartered in Emeryville, California.
• Skild AI: Robotics company building an omni-bodied robot foundation model, the Skild Brain, to control any robot for any task.
• 1X Technologies: Humanoid robotics company, founded as Halodi Robotics, building home robots and a world model for robot learning.
• Figure AI: Humanoid robot company founded by Brett Adcock, developer of the Figure robots and the Helix vision-language-action model.
• Inception (G42): G42's AI product and research company, known for the Jais Arabic-centric open language models.
• Wayve: London autonomous driving company using end-to-end learning, known for the GAIA generative driving world models.
• LightOn: French enterprise generative AI company, co-developer of ModernBERT and maker of late-interaction retrieval models.
• DeepL: German AI research company known for the DeepL Translator and its own translation-focused language models.
• H Company: Paris startup building computer-use agents and the open Holo models that navigate user interfaces.
• Aleph Alpha: German company building sovereign LLMs for enterprises and governments, including Pharia models and tokenizer-free T-FREE research.
• Twelve Labs: Video-native AI company with offices in San Francisco and Seoul, building the Marengo and Pegasus video models.
• Synthesia: London AI video company that researches neural rendering and synthetic avatars for business video.
• Lightricks: Jerusalem creative-app company behind Facetune and the open LTX-Video real-time video generation models.
• Decart: Real-time generative video lab, known for Oasis, a playable AI-generated game world model.
• Sesame: Conversational voice company whose Conversational Speech Model aims to cross the uncanny valley of AI voice.
• Hume AI: Voice AI company grounded in emotion science, known for the Empathic Voice Interface and Octave text-to-speech.
• Cartesia: Real-time voice AI company pioneering new model architectures, maker of the Sonic text-to-speech models.
• ElevenLabs: AI audio research and product company known for natural-sounding speech synthesis and voice agent models.
• Suno: Music generation company whose models create songs with vocals from prompts; earlier released the open Bark audio model.
• Ideogram: Maker of the Ideogram text-to-image models, noted for rendering legible text inside generated images.
• Genmo: Video generation lab that released Mochi 1, an open-source text-to-video model.
• Luma AI: Generative video company behind Dream Machine and the Ray video models, with research on pretraining such as Inductive Moment Matching.
• Runway: New York generative video company behind the Gen-3 and Gen-4 video models and the GWM-1 world model.
• Midjourney: Self-described independent research lab led by David Holz, known for the Midjourney text-to-image models.
• Cleanlab: Data-centric AI company built on confident learning, a method for finding label errors in datasets.
• DatologyAI: Data curation company that researches how selecting and synthesizing training data makes models better and cheaper.
• Snorkel AI: Data-centric AI company spun out of the Stanford AI Lab, originating from the Snorkel weak supervision project.
• Jina AI: Search AI company founded by Han Xiao, known for open embedding, reranker and reader models.
• Voyage AI: Embedding and reranking model company for retrieval, acquired by MongoDB in 2025.
• Nomic AI: Company behind GPT4All for local LLMs and the reproducible Nomic Embed text embedding models.
• Unsloth AI: Open-source toolkit maker for fast, memory-efficient fine-tuning and local running of LLMs, with many quantized models.
• Periodic Labs: AI research company building AI scientists and autonomous laboratories for the physical sciences.
• Harmonic: Company building mathematical superintelligence; its Aristotle system reached gold-medal level at the 2025 IMO.
• Zyphra: San Francisco lab training open multimodal models, known for Zamba SSM-hybrid language models and Zonos text-to-speech.
• Arcee AI: US company releasing open-weight models such as AFM and Trinity, and creator of the MergeKit model-merging toolkit.
• Prime Intellect: San Francisco company training models over decentralized compute, known for the INTELLECT series and RL environments.
• Cerebras Systems: Maker of wafer-scale AI chips whose research team released open Cerebras-GPT and BTLM language models.
• Together AI: AI cloud company co-founded by researchers including Tri Dao, with research such as RedPajama and Mixture-of-Agents.
• Answer.AI: R&D lab from Jeremy Howard and Eric Ries turning foundational research into practical tools, such as FSDP+QLoRA and ModernBERT.
• Inception Labs: AI research and product company building diffusion-based large language models, starting with Mercury.
• Deep Cogito: San Francisco lab releasing open Cogito models trained with iterated distillation and amplification.
• Reflection AI: Lab founded by former Google DeepMind researchers Misha Laskin and Ioannis Antonoglou to build open frontier models.
• Perplexity: AI answer engine company that also releases open models such as the pplx-embed embedding family.
• Writer: Enterprise generative AI platform company that trains its own Palmyra family of large language models.
• Contextual AI: Enterprise RAG company founded by RAG co-author Douwe Kiela, known for RAG 2.0 and the KTO alignment method.
• Imbue: AI lab founded in 2021 that trained a 70B model from scratch and builds tools that put software under users' control.
• Cognition: San Francisco company behind Devin, an autonomous AI software engineer, and its own SWE agent models.
• Poolside: AI company developing foundation models for software development, including the open-weight Laguna coding models.
• Magic: AGI startup building code models with ultra-long context, including 100M-token context windows and the HashHop eval.
• Essential AI: San Francisco lab led by Ashish Vaswani, publishing open pretraining research, datasets and the Rnj-1 model.
• Adept AI Labs: Agent startup that trained models to act in software, releasing the open Persimmon-8B and Fuyu-8B multimodal models.
• Inflection AI: Public benefit corporation that built Pi, a personal AI assistant powered by the Inflection AI model stack.
• Character.AI: Consumer chatbot company founded by LaMDA developers Noam Shazeer and Daniel de Freitas, training its own conversational models.
• Scale Labs (Scale AI): Scale AI's research arm for evaluation and alignment, co-creator of Humanity's Last Exam and SWE-Bench Pro.
• Waymo Research: Waymo's autonomous driving research group, publisher of the Waymo Open Dataset and end-to-end driving models such as EMMA.
• Toyota Research Institute: Toyota's research subsidiary for AI, robotics, automated driving and materials, known for diffusion-policy robot learning.
• Yandex Research: Yandex's research team for machine learning, known for CatBoost gradient boosting and the open YaLM-100B model.
• ServiceNow AI Research: ServiceNow's AI research group, co-steward of the BigCode StarCoder project and maker of the Apriel models.
• Snap Research: Snap's research group, known for efficient generative image models that run on mobile devices.
• J.P. Morgan AI Research: JPMorgan Chase's AI research group studying AI agents, finance foundation models, synthetic data and document AI.
• Bloomberg AI Engineering: Bloomberg's AI group, known for BloombergGPT, a 50-billion-parameter language model trained for finance.
• Spotify Research: Spotify's research team working on audio intelligence, recommendation and search, known for the Podcast Dataset and Basic Pitch.
• Netflix Research: Netflix's research group publishing on machine learning, recommendation and video encoding.
• Qualcomm AI Research: Qualcomm's AI research group, focused on efficient on-device AI such as quantization and model compression.
• Intel Labs: Intel's research organization; its Intelligent Systems Lab created the MiDaS monocular depth models and DPT.
• Adobe Research: Adobe's research lab for graphics, imaging and generative AI, including 3D reconstruction and fast diffusion distillation.
• Samsung Advanced Institute of Technology (SAIT): Samsung's advanced research institute; its SAIL Montreal AI lab released the Tiny Recursive Model for puzzle reasoning.
• Samsung Research: Samsung Electronics' R&D hub, running AI centers in Seoul, Mountain View, Toronto, Cambridge, Warsaw, New York and Montreal.
• Sarvam AI: Bengaluru company building language models for Indian languages, part of India's sovereign LLM programme.
• Upstage: South Korean AI company known for the Solar language models and depth up-scaling, plus document AI.
• SB Intuitions: SoftBank subsidiary developing homegrown Japanese large language models under the Sarashina name.
• Skywork AI (Kunlun Tech): Kunlun Tech's AI unit releasing the open Skywork language, reasoning and reward models.
• ModelBest (OpenBMB): Tsinghua-incubated Beijing company best known for the open MiniCPM small language and on-device multimodal models.
• Megvii: Beijing computer vision company behind Face++, with research on efficient networks such as ShuffleNet and the YOLOX detector.
• RedNote hi lab: Xiaohongshu's AI lab, releasing the open dots.llm1 language model and dots.ocr document parser.
• Alibaba DAMO Academy: Alibaba's research academy, whose language technology lab builds VideoLLaMA and the SeaLLMs Southeast Asian language models.
• inclusionAI (Ant Group): Ant Group's open model lab behind the Ling language models and Ring reasoning models.
• Meituan LongCat: Meituan's LongCat team, which releases the open LongCat-Flash mixture-of-experts language models.
• Kling AI (Kuaishou): Kuaishou's generative video unit behind the Kling video model, plus open work such as LivePortrait and Kolors.
• Xiaomi MiMo: Xiaomi's large model team, releasing the open MiMo reasoning and MiMo-VL vision-language models.
• Baichuan AI: Beijing AI company, one of China's AI Tigers, known for the open Baichuan language models and medical models.
• StepFun: Shanghai AI company, one of China's AI Tigers, releasing open speech, video and multimodal reasoning models.
• iFLYTEK: Chinese speech and language AI company, known for voice recognition and, with HIT, the Chinese BERT-wwm and MacBERT models.
• SenseTime: Hong Kong-listed AI company known for computer vision research and the SenseNova foundation model series.
• Huawei Noah's Ark Lab: Huawei's AI research center, known for PanGu language models and efficient models such as TinyBERT and GhostNet.
• Preferred Networks: Japanese deep learning company that created Chainer and Optuna and develops the PLaMo Japanese language models and its own AI chips.
• Sony AI: Sony Group's AI research subsidiary, known for the GT Sophy racing agent and the FHIBE fairness benchmark dataset.
• Kakao AI (Kanana): Kakao's AI effort behind the Kanana bilingual Korean-English models, building on Kakao Brain's earlier KoGPT.
• NAVER Cloud AI (CLOVA): NAVER's AI research organization behind the HyperCLOVA X Korean language models and vision work such as CutMix and Donut.
• LG AI Research: LG Group's AI research institute and developer of the EXAONE family of Korean-English language and reasoning models.
• Institute for Artificial Intelligence, Peking University: Peking University institute announced on 27 April 2019 to coordinate university-wide AI research and training.
• Institut national pour l'évaluation et la sécurité de l'intelligence artificielle (INESIA): France's National Institute for AI Evaluation and Security, created by the French government on 31 January 2025.
• European AI Office: European Commission body that implements the EU AI Act, especially rules for general-purpose AI models.
• Singapore AI Safety Institute (Digital Trust Centre): Singapore's AI safety institute: the NTU Digital Trust Centre, founded in 2022 and designated the AISI in 2024, working with IMDA.
• Korea AI Safety Institute: South Korea's AI safety institute, founded in November 2024 under ETRI and based in Bundang, Seongnam.
• Canadian Artificial Intelligence Safety Institute (CAISI): Canada's AI safety institute, founded in November 2024 and led by ISED with NRC and CIFAR research partners.
• Japan AI Safety Institute (J-AISI): Japan's AI safety institute, launched on 14 February 2024 within IPA to develop AI safety evaluation methods and standards.
• Centre for the Governance of AI (GovAI): AI governance think tank founded at Yale in 2016, now an independent non-profit with offices in London and Washington, DC.
• Center for Human-Compatible Artificial Intelligence (CHAI): UC Berkeley AI safety research center founded in 2016 by a group led by Stuart Russell.
• Distributed AI Research Institute (DAIR): Independent, community-rooted AI research institute founded by Timnit Gebru in December 2021.
• Software Heritage: Non-profit archive of publicly available source code, launched by Inria in 2016 with UNESCO support.
• Common Crawl Foundation: 501(c)(3) nonprofit that crawls the web monthly and publishes a free open archive of web data collected since 2008.
• AI Now Institute: Independent research institute founded in 2017 producing policy research on the social impact of AI and industry power.
• Transluce: Independent nonprofit research lab building open tools for public oversight and understanding of AI systems.
• Alignment Research Center (ARC): Berkeley non-profit founded by Paul Christiano in 2021, working on theory for mechanistic explanations of neural networks.
• Machine Intelligence Research Institute (MIRI): Berkeley nonprofit founded in 2000 as the Singularity Institute, focused on existential risk from superintelligent AI.
• FAR.AI: AI security research and adoption non-profit, known for adversarial attacks on superhuman Go AIs and the tuned lens.
• Redwood Research: Nonprofit AI safety and security research organization that pioneered the AI control agenda.
• Apollo Research: AI safety organization focused on evaluating and monitoring scheming and deception in frontier AI systems.
• Center for AI Safety (CAIS): San Francisco nonprofit doing technical AI safety research and field-building; organized the 2023 Statement on AI Risk.
• AI Singapore: Singapore's national AI programme launched in May 2017, which develops the SEA-LION Southeast Asian language models.
• Technology Innovation Institute (TII): Abu Dhabi government-funded research institute, part of ATRC, that develops the Falcon open language models.
• Mohamed bin Zayed University of Artificial Intelligence (MBZUAI): Graduate research university dedicated to AI in Masdar City, Abu Dhabi, licensed in 2020.
• Wadhwani Institute for Artificial Intelligence (Wadhwani AI): Independent non-profit founded in 2018 building AI for underserved communities in health, agriculture and education.
• AI4Bharat: IIT Madras lab building open datasets, tools and models for Indian languages.
• Matsuo-Iwasawa Laboratory, University of Tokyo: University of Tokyo deep learning lab working on world models, robotics and LLMs, with many spin-off startups.
• National Institute of Informatics (NII) and LLM-jp: Japan's national informatics institute, established in 2000, whose LLM R&D center runs the open LLM-jp project.
• RIKEN Center for Advanced Intelligence Project (AIP): RIKEN AI research center launched in 2016, covering mathematical foundations of ML, scientific AI and societal AI.
• Beijing Institute for General Artificial Intelligence (BIGAI): Beijing AGI research institute founded in 2020 by Song-Chun Zhu with Beijing municipal government support.
• Institute of Automation, Chinese Academy of Sciences (CASIA): Chinese Academy of Sciences institute founded in 1956 for intelligence science, pattern recognition and robotics.
• Tsinghua NLP Lab (THUNLP): Tsinghua natural language processing lab known for knowledge-enhanced language models and the MiniCPM small LLMs.
• Institute for AI Industry Research (AIR), Tsinghua University: Tsinghua institute for applied AI research in health, transportation and IoT, led by Dean Zhang Ya-Qin.
• IVADO: Quebec's largest interdisciplinary AI research, training and knowledge transfer consortium, led by Université de Montréal.
• Alberta Machine Intelligence Institute (Amii): Independent non-profit AI institute in Edmonton, begun in 2002 as AICML and one of three Pan-Canadian AI Strategy institutes.
• University of Toronto Machine Learning Group: Machine learning faculty group in Toronto's Department of Computer Science, home of the AlexNet and dropout papers.
• Swiss AI Initiative: Swiss open foundation model effort started in December 2023 by ETH Zurich, EPFL and CSCS, which released the Apertus LLM.
• Dalle Molle Institute for Artificial Intelligence (IDSIA): Swiss AI institute in Lugano founded in 1988, known for early deep learning and recurrent network research.
• German Research Center for Artificial Intelligence (DFKI): German AI research center founded in 1988 as a non-profit public-private partnership, with sites across Germany.
• Tübingen AI Center: Tübingen machine learning center of about 25 research groups, funded by the German federal and Baden-Württemberg governments.
• EPFL Machine Learning and Optimization Laboratory (MLO): EPFL lab on optimization and distributed machine learning, known for sent2vec and the MEDITRON medical LLMs.
• Inria (National Institute for Research in Digital Science and Technology): French national research institute for computer science and applied mathematics, where scikit-learn and Software Heritage began.
• European Laboratory for Learning and Intelligent Systems (ELLIS): Pan-European AI network of excellence founded in 2018, with institutes and units at research institutions across Europe.
• Amsterdam Machine Learning Lab (AMLab): University of Amsterdam lab on deep generative models, probabilistic ML and geometric deep learning.
• The Alan Turing Institute: UK national institute for data science and AI, an independent charity based at the British Library, founded in 2015.
• EdinburghNLP: University of Edinburgh NLP group, known for machine translation work including subword (BPE) units and the Nematus toolkit.
• Gatsby Computational Neuroscience Unit: UCL unit set up by the Gatsby Charitable Foundation in 1998 for theoretical neuroscience and machine learning; Geoffrey Hinton was founding director.
• Cambridge Machine Learning Group (MLG): Machine learning group in the Cambridge Department of Engineering, known for probabilistic and Bayesian machine learning.
• Visual Geometry Group (VGG), University of Oxford: Oxford computer vision group known for the VGG convolutional networks and the VGGFace2 dataset.
• Oxford Applied and Theoretical Machine Learning Group (OATML): Oxford Computer Science group led by Yarin Gal, working on Bayesian deep learning, uncertainty and AI safety.
• MIT Data Systems Group (DSG): MIT CSAIL group researching data management, from query processing to learned database systems such as SageDB.
• UCLA NLP: UCLA NLP group, known for the VisualBERT vision-language model and the MathVista benchmark.
• Machine Learning Center at Georgia Tech (ML@GT): Interdisciplinary machine learning research center at Georgia Tech, founded in 2016.
• CMU Robotics Institute: CMU robotics department founded in 1979, known for computer vision, autonomous vehicles and the first robotics PhD program.
• CMU Language Technologies Institute (LTI): CMU institute for language technologies, begun in 1986 as the Center for Machine Translation and renamed LTI in 1996.
• NSF AI Institute for Foundations of Machine Learning (IFML): NSF AI institute led by UT Austin, designated in 2020, working on foundational machine learning theory and tools.
• UW Database Group: University of Washington group on data management systems, cloud services, probabilistic databases and data pricing.
• UW NLP: University of Washington natural language processing group behind ELMo, Self-Instruct and QLoRA.
• Columbia Data Science Institute: University-wide data science institute at Columbia, founded in 2012 within Columbia Engineering.
• CILVR Lab (Computational Intelligence, Learning, Vision, and Robotics): NYU machine learning and vision lab founded in 2009 by Yann LeCun and Rob Fergus.
• NYU Center for Data Science: NYU's university-wide data science center, founded in 2013 with Yann LeCun as founding director.
• Princeton NLP Group: Princeton natural language processing group behind SimCSE, SWE-bench and Tree of Thoughts.
• Princeton Language and Intelligence (PLI): Princeton university-wide initiative on large AI models and their use across the sciences and humanities.
• Kempner Institute for the Study of Natural and Artificial Intelligence: Harvard institute studying intelligence in brains and machines, funded by a Chan Zuckerberg gift in 2021.
• MIT-IBM Computing Research Lab (formerly MIT-IBM Watson AI Lab): MIT and IBM joint research lab, evolved from the MIT-IBM Watson AI Lab that began in 2017.
• Stanford DAWN: Five-year Stanford project (2018 to 2022) on tools to democratize ML, including DAWNBench, Snorkel and ColBERT.
• Hazy Research: Chris Re's Stanford group on ML systems, known for FlashAttention, Snorkel weak supervision and structured state space models.
• Stanford NLP Group: Stanford natural language processing group behind GloVe, SQuAD, CoreNLP, Stanza and DSPy.
• Stanford Artificial Intelligence Laboratory (SAIL): Stanford's umbrella AI lab for research and teaching across machine learning, vision, NLP and robotics since 1963.

Open image-text datasets (LAION-400M, LAION-5B) that trained Stable Diffusion and OpenCLIP.

And 63 more labs: Kyutai, TUM Database Systems, CWI Database Architectures and othersShow all

• Kyutai: Open-science lab funded by Xavier Niel and partners: Moshi real-time voice and Hibiki live translation.
• TUM Database Systems: Compiled query execution in main-memory and disk-based systems (HyPer, Umbra), from Thomas Neumann's group.
• CWI Database Architectures: Column stores and vectorized execution: MonetDB, X100 (Vectorwise) and the research that became DuckDB.
• DuckLabs (DuckDB Labs): The company of DuckDB's creators, the in-process analytics database born at CWI; Amazon agreed to buy it in 2026.
• METR: Evaluating dangerous capabilities: the time-horizon measure of how long a task AI agents can finish.
• Epoch AI: Data on AI's trajectory: training compute trends, data limits, and the FrontierMath benchmark.
• US Center for AI Standards and Innovation (CAISI): NIST's AI testing center, founded as the US AI Safety Institute and renamed CAISI in 2025; evaluates frontier and foreign models.
• UK AI Security Institute: The first state AI safety institute (renamed Security in 2025): pre-deployment testing and the Inspect eval framework.
• ETH AI Center: ETH Zurich's cross-department AI hub and a driver of the Swiss AI Initiative's open Apertus model.
• Max Planck Institute for Intelligent Systems: Causal inference and kernel methods (Scholkopf), and the SMPL body model used across graphics and vision.
• Shanghai AI Laboratory: The InternLM and InternVL open models and the OpenCompass evaluation platform.
• BAAI (Beijing Academy of AI): The BGE embedding models used across open retrieval stacks, the Emu multimodal models and FlagOpen.
• Tsinghua KEG: Tsinghua's knowledge engineering group: AMiner, the GLM and ChatGLM models (with Zhipu), CogVLM and AgentBench.
• Vector Institute: Toronto's independent AI institute, home to deep learning researchers and work such as neural ODEs.
• Mila: Yoshua Bengio's institute: neural language models, attention for translation, GANs and Theano.
• CMU Database Group: Database systems research and teaching in the open: self-driving databases, NoisePage and the Database of Databases.
• CMU Machine Learning Department: The world's first machine learning department: Libratus poker, Transformer-XL, XLNet and Mamba.
• MIT CSAIL: MIT's largest lab: robotics, the Julia language, adversarial robustness and the lottery ticket hypothesis.
• Berkeley Sky Computing Lab: The systems lab after AMPLab (Spark) and RISELab (Ray): vLLM, SkyPilot and Chatbot Arena.
• Berkeley AI Research (BAIR): Diffusion models (DDPM), NeRF, robot learning, and open chat models such as Vicuna.
• Stanford HAI and CRFM: The AI Index, the term foundation models, the HELM benchmark and Alpaca, from Stanford's AI institute.
• EleutherAI: Grassroots open research: The Pile, GPT-NeoX and Pythia, and the LM Evaluation Harness everyone uses.
• Ai2 (Allen Institute for AI): Fully open models with their data and training code: OLMo, Tulu, Molmo, and the Dolma corpus.
• Isomorphic Labs: AI-first drug discovery spun out of DeepMind; co-developed AlphaFold 3.
• World Labs: Fei-Fei Li's spatial intelligence company: models that generate explorable 3D worlds (Marble).
• Physical Intelligence: General-purpose robot foundation models: the pi-zero vision-language-action family, open-sourced as openpi.
• Nous Research: The open Hermes fine-tunes, YaRN context extension, and decentralized training over the internet.
• Liquid AI: Liquid foundation models (LFM) for devices, built on liquid neural network research from MIT CSAIL.
• Stability AI: Stable Diffusion, the open image model release that opened up generative imaging in 2022.
• Black Forest Labs: The FLUX image models, from the researchers behind latent diffusion and Stable Diffusion.
• Tencent Hy (Hunyuan): Tencent's Hunyuan models, with open releases in MoE language, video (HunyuanVideo) and 3D generation.
• Baidu (ERNIE): The ERNIE models and PaddlePaddle, and earlier Deep Speech 2 from its Silicon Valley lab.
• ByteDance Seed: ByteDance's model team behind Doubao, the UI-TARS computer-use agents and Seedance video generation.
• Hugging Face: The Hub for open models and datasets and the Transformers library; also BLOOM, StarCoder, FineWeb and SmolLM.
• Snowflake AI Research: Arctic: open embedding models, the Arctic LLM, SwiftKV inference and text-to-SQL research.
• Databricks Mosaic Research: Open models (MPT, DBRX) and training at scale, from the MosaicML team Databricks bought in 2023; Spark lineage.
• Salesforce AI Research: BLIP vision-language models, CodeT5, and xLAM action models for agents.
• IBM Research: Deep Blue and Watson, relational databases and quantum computing, and today the open Granite models.
• Amazon AGI and Amazon Science: The Amazon Nova models and Nova Act agents, plus Chronos forecasting models from Amazon Science.
• Apple Machine Learning Research: On-device models for Apple Intelligence, the MLX framework for Apple silicon, and studies of reasoning limits.
• NVIDIA Research: StyleGAN, Megatron-LM for training giant models, Instant NeRF, and the open Nemotron models.
• Safe Superintelligence (SSI): Ilya Sutskever's lab with one goal and one product, a safe superintelligence; it has published nothing yet.
• Thinking Machines Lab: Mira Murati's lab: Tinker, a fine-tuning API, and open research notes on reproducible inference.
• Sakana AI: Nature-inspired methods: evolutionary model merging and The AI Scientist, from a Transformer co-author.
• Reka: Natively multimodal models (Core, Flash, Edge) from former DeepMind and Meta researchers.
• AI21 Labs: Jamba, the first production-scale hybrid of Transformer and Mamba layers.
• Cohere and Cohere Labs: Enterprise models (Command, Embed, Rerank) and Cohere Labs' open multilingual Aya research.
• 01.AI: Kai-Fu Lee's lab behind the open Yi models, now focused on enterprise AI products.
• MiniMax: Lightning attention for million-token context in open models, plus Hailuo video and speech models.
• Moonshot AI: The Kimi models: long context early, then the open-weight Kimi K2 trained with the Muon optimizer.
• Z.ai (Zhipu AI): The GLM models, spun out of Tsinghua University's KEG lab; GLM-4.5 and later are open weights.
• Alibaba Qwen: The Qwen family, the most downloaded open-weight models, from Alibaba Cloud's Tongyi lab.
• DeepSeek: Frontier-class open-weight models trained cheaply: DeepSeek-V3's efficient MoE and R1's open reasoning recipe.
• Mistral AI: Europe's leading model maker: efficient open-weight models, Mixtral's mixture of experts, and Le Chat.
• xAI (SpaceXAI): The Grok models and the Colossus supercomputer in Memphis; now a division of SpaceX.
• Microsoft AI: Copilot for consumers and Microsoft's own MAI models, led by Mustafa Suleyman since 2024.
• Microsoft Research: ResNet, LoRA, DeepSpeed and the small Phi models, from a lab spanning systems, theory and AI.
• Meta Superintelligence Labs: Meta's frontier model effort, formed in 2025 under Alexandr Wang; it builds the Muse models and now houses FAIR.
• Meta FAIR: PyTorch, the open Llama models, Segment Anything and self-supervised vision (DINO).
• Google Research: The Transformer, BERT and T5, plus MapReduce and TensorFlow: much of the base of modern AI.
• Google DeepMind: AlphaGo and AlphaFold, deep reinforcement learning, and Google's Gemini models.
• Anthropic: The Claude models, Constitutional AI and interpretability research on what happens inside models.
• OpenAI: The GPT models and ChatGPT, and the scaling results that set off the large language model race.