Model organizations

OpenAI develops foundation models for general-purpose reasoning, generation, and agentic workflows.

Alibaba develops the Qwen family and related multimodal and coding models.

Google develops the Gemini family and researches language, vision, audio, and multimodal systems.

Anthropic is an AI safety and research company building the Claude family of models.

Moonshot AI develops the Kimi family of language models for long-context reasoning and general-purpose use.

Zhipu AI develops the GLM family of foundation models for language, reasoning, coding, and multimodal applications.

DeepSeek develops foundation models focused on reasoning, coding, and efficient inference.

Mistral AI builds efficient foundation models for open and commercial AI applications.

MiniMax develops foundation models for language, multimodal generation, and interactive AI applications.

NVIDIA develops the Nemotron family and supporting AI research and infrastructure for model training and inference.

Meta AI develops open model families and research systems for language, vision, and multimodal applications.

xAI develops the Grok family of models for reasoning, conversation, and multimodal use cases.

Cohere builds enterprise-oriented language models for search, retrieval, generation, and multilingual applications.

Xiaomi develops the MiMo family of language models for reasoning and general-purpose AI applications.

Tencent develops AI models including the Hunyuan and Hy families for language and multimodal applications.

ByteDance Seed develops the Seed family of foundation models for language and multimodal AI.

Thinking Machines is represented in this catalog through the Ling family of language models.

StepFun develops the Step family of general-purpose language models.

InclusionAI is represented in this catalog through the Ling and Ring families of language models.

Poolside is represented in this catalog through the Laguna family of code-focused language models.

Aion Labs is represented in this catalog through the Aion family of language models.

Kwaipilot is represented in this catalog through the Kat family of language models.

Sakana AI researches adaptive and efficient AI systems, including the Fugu family represented in this catalog.

Microsoft develops AI models including the MAI, Phi, and WizardLM families for general-purpose and specialized workloads.

Perplexity develops answer and search-oriented AI systems combining language models with retrieval.

DeepReinforce is represented in this catalog through the Ornith family of language models.

Nex Agi is represented in this catalog through the Nex family of language models.

Meituan is represented in this catalog through the Longcat family of language models.

IBM develops the Granite family of enterprise-focused foundation models for language, code, and other AI workloads.

Arcee AI develops efficient language models, including the Trinity and Virtuoso families.

Perceptron is represented in this catalog through the Perceptron family of language models.

Amazon develops the Nova family of foundation models for text, vision, speech, and multimodal applications.

Reka AI develops multimodal foundation models for language, vision, and long-context applications.

Sarvam AI develops language models and AI systems focused on Indian languages and applications.

Inception is represented in this catalog through the Mercury family of language models.

Upstage develops the Solar family of language models and document AI systems for enterprise use.

Writer develops enterprise generative AI systems and the Palmyra family of language models.

TheDrummer is represented in this catalog through the Cydonia, Rocinante, Skyfall, and Unslopnemo model families.

Relace develops code-focused language models and developer tools, including the Relace family.

Kuaishou develops the Kling family of multimodal image and video generation models.

Nous Research develops open language models, including the Hermes family, with a focus on accessible research and customization.

AI2 develops open research models, including the OLMo family, for transparent and reproducible AI research.

DeepCogito is represented in this catalog through the Cogito family of reasoning models.

Runway develops generative models and creative tools for image and video production.

Morph develops code-focused language models and tools for fast software development.

Baidu develops foundation models including the ERNIE family for language, reasoning, and multimodal applications.

AI21 develops foundation models for language understanding and generation, including the Jamba family.

ByteDance develops AI models for generative and interactive applications, including the UI family represented here.

Cognitive Computations is represented in this catalog through the Dolphin family of language models.

Sao10k is represented in this catalog through the L3 family of language models.

Anthracite Org is represented in this catalog through the Magnum family of language models.

Undi95 is represented in this catalog through the Remm family of language models.

Mancer is represented in this catalog through the Weaver family of language models.

Gryphe is represented in this catalog through the MythoMax family of language models.