LLM AtlasLLM AtlasSearch models

Providers

Provider directory

27 providers

Compare the major labs and vendors building LLMs and multimodal systems, with a quick read on catalog depth, access mix, freshness, and how much of each lineup is backed by high-confidence sourcing.

Last verified timestamps and source confidence are attached to every provider record. Use the directory for breadth, then open provider pages for source links, model freshness, and full-profile coverage details.

API providers
20

Labs with directly tracked API-accessible models.

High-confidence profiles
27

Providers whose catalog records are backed by stronger primary-source coverage.

Last verified
2026-03-29

Coverage spans API and open-weight providers, including vendors tracked primarily for current market visibility.

Open-weight providers
14

Organizations with notable downloadable or self-hosted model families.

Tracked models
270

Combined provider coverage across current, legacy, and newly added frontier vendors.

OpenAI

Frontier API model provider spanning reasoning, realtime, image, audio, and open-weight releases.

31 models
Headquarters
San Francisco, CA
Catalog mix
29 API / 2 open-weight
31 full-profile entries · high confidenceView provider

Anthropic

Provider focused on Claude reasoning models, safety, long-context workflows, and enterprise deployments.

19 models
Headquarters
San Francisco, CA
Catalog mix
19 API / 0 open-weight
19 full-profile entries · high confidenceView provider

Google DeepMind

Gemini model provider with strong multimodal, long-context, and agentic capabilities.

16 models
Headquarters
Mountain View, CA
Catalog mix
16 API / 0 open-weight
16 full-profile entries · high confidenceView provider

Meta

Publisher of the Llama family and related safety models distributed as open-weight releases.

17 models
Headquarters
Menlo Park, CA
Catalog mix
0 API / 17 open-weight
17 full-profile entries · high confidenceView provider

Mistral AI

Commercial and open-model provider covering generalist, coding, reasoning, OCR, and audio model lines.

25 models
Headquarters
Paris, France
Catalog mix
14 API / 11 open-weight
25 full-profile entries · high confidenceView provider

Cohere

Enterprise-oriented provider with Command model families optimized for RAG and agents.

8 models
Headquarters
Toronto, Canada
Catalog mix
8 API / 0 open-weight
8 full-profile entries · high confidenceView provider

xAI

Provider of Grok reasoning and agentic API models.

5 models
Headquarters
Palo Alto, CA
Catalog mix
5 API / 0 open-weight
5 full-profile entries · high confidenceView provider

DeepSeek

Lab publishing DeepSeek reasoning, coding, OCR, multimodal, and open model releases.

14 models
Headquarters
Hangzhou, China
Catalog mix
0 API / 14 open-weight
14 full-profile entries · high confidenceView provider

Alibaba Qwen

Publisher of the Qwen family spanning generalist, coder, vision, and mixture-of-experts models.

18 models
Headquarters
Hangzhou, China
Catalog mix
0 API / 18 open-weight
18 full-profile entries · high confidenceView provider

Microsoft

Publisher of the Phi family of compact language, reasoning, and multimodal models.

17 models
Headquarters
Redmond, WA
Catalog mix
0 API / 17 open-weight
17 full-profile entries · high confidenceView provider

AI21 Labs

Publisher of the Jamba family of long-context open and enterprise models.

6 models
Headquarters
Tel Aviv, Israel
Catalog mix
0 API / 6 open-weight
6 full-profile entries · high confidenceView provider

MiniMax

Provider of the MiniMax M-series, speech, video, music, and agent APIs for multimodal application builders.

12 models
Headquarters
Shanghai, China
Catalog mix
12 API / 0 open-weight
12 full-profile entries · high confidenceView provider

Z.AI

Provider of the GLM family spanning reasoning, coding, multimodal, image, video, audio, and agent APIs.

11 models
Headquarters
Beijing, China
Catalog mix
11 API / 0 open-weight
11 full-profile entries · high confidenceView provider

Moonshot AI / Kimi

Provider of Kimi API models focused on long-context, coding, agentic reasoning, and web-connected workflows.

4 models
Headquarters
Beijing, China
Catalog mix
4 API / 0 open-weight
4 full-profile entries · high confidenceView provider

ByteDance / Doubao

ByteDance's Doubao family distributed through Volcengine Ark across text, vision, image, video, speech, and agent services.

4 models
Headquarters
Beijing, China
Catalog mix
4 API / 0 open-weight
4 full-profile entries · high confidenceView provider

Tencent / Hunyuan

Tencent HY provider spanning text, reasoning, translation, vision, image, video, and 3D model lines.

7 models
Headquarters
Shenzhen, China
Catalog mix
7 API / 0 open-weight
7 full-profile entries · high confidenceView provider

Baidu / ERNIE

Baidu's ERNIE family is delivered through Qianfan for text, reasoning, multimodal, and enterprise agent workflows.

5 models
Headquarters
Beijing, China
Catalog mix
5 API / 0 open-weight
5 full-profile entries · high confidenceView provider

Amazon Web Services

AWS Bedrock provider offering Nova models and third-party model access for enterprise cloud workloads.

3 models
Headquarters
Seattle, WA
Catalog mix
3 API / 0 open-weight
3 full-profile entries · high confidenceView provider

IBM

Provider of Granite open-source models through watsonx for enterprise AI, code, and data tasks.

3 models
Headquarters
Armonk, NY
Catalog mix
3 API / 3 open-weight
0 full-profile entries · high confidenceView provider

Stability AI

Provider of Stable Diffusion and other generative models for image creation and editing.

2 models
Headquarters
London, UK
Catalog mix
2 API / 2 open-weight
0 full-profile entries · high confidenceView provider

Black Forest Labs

Creator of FLUX image generation models, offering both open-weight and API-based text-to-image systems.

15 models
Headquarters
Freiburg, Germany
Catalog mix
15 API / 12 open-weight
3 full-profile entries · high confidenceView provider

NVIDIA

Publisher of the Nemotron family of open enterprise models, embedding models, and reasoning-optimized LLMs.

7 models
Headquarters
Santa Clara, CA
Catalog mix
7 API / 7 open-weight
0 full-profile entries · high confidenceView provider

Perplexity

Search-augmented AI platform offering Sonar models, agent API, search API, and embeddings for real-time research and Q&A.

6 models
Headquarters
San Francisco, CA
Catalog mix
6 API / 0 open-weight
6 full-profile entries · high confidenceView provider

Morph

Specialized code-editing and search models for AI coding agents. Fast Apply (10,500 tok/s), WarpGrep code search, and context compaction.

3 models
Headquarters
San Francisco, CA
Catalog mix
3 API / 0 open-weight
3 full-profile entries · high confidenceView provider

Xiaomi

Publisher of the MiMo family of open-weight models including MiMo-V2-Flash (309B MoE), MiMo-Audio, MiMo-VL, and MiMo-Embodied.

3 models
Headquarters
Beijing, China
Catalog mix
1 API / 3 open-weight
2 full-profile entries · high confidenceView provider

StepFun

Chinese AI lab publishing open-weight models including Step-3.5-Flash (199B), Step3-VL vision-language models, and Step-Audio models.

4 models
Headquarters
Beijing, China
Catalog mix
0 API / 4 open-weight
4 full-profile entries · high confidenceView provider

Liquid AI

Publisher of LFM2 hybrid models designed for on-device deployment. MoE architectures from 350M to 24B parameters with edge-optimized inference.

5 models
Headquarters
Boston, MA
Catalog mix
0 API / 5 open-weight
5 full-profile entries · high confidenceView provider