14 models US lab that created ChatGPT; it also makes the GPT Image models and the Codex coding agent.
For the public: ChatGPT
Text 4 Coding 4 Image 9 Decision 1 Open source 1
See its models 4 models US lab focused on AI safety; its models are called Claude (Haiku, Sonnet, Opus and Fable).
For the public: Claude
Text 4 Coding 4
See its models 18 models Its assistant is Gemini; it also creates images (Nano Banana, Imagen) and video (Veo) and publishes the open Gemma models.
For the public: Gemini
Text 3 Coding 3 Image 9 Video 4 Decision 2 Open source 3
See its models 6 models Elon Musk's AI company; its assistant Grok is on the web, in apps and inside X.
For the public: Grok
Text 1 Coding 1 Image 3 Video 2
See its models 3 models The company behind Facebook, Instagram and WhatsApp; its Muse Spark models power the Meta AI assistant, and it previously published the open Llama models.
For the public: Meta AI
Text 2 Coding 2 Image 1 Open source 1
See its models 2 models French lab (European Union) that publishes open and closed models.
For the public: Le Chat
Text 2 Coding 2 Open source 1
See its models 26 models The Chinese e-commerce giant; its Qwen models are among the most used open-source models in the world.
For the public: Qwen Chat
Text 4 Coding 4 Image 16 Video 5 Decision 2 Open source 12
See its models HappyHorse-1.0 (video) HappyHorse-1.1 (video) Qwen Image (image, open source) Qwen Image 2.0 (2026-03-03) (image) Qwen Image 2.0 Pro (2026-04-22) (image) Qwen Image Max 2512 (image, open source) Qwen Image Plus 2601 (image) Qwen-Image-2.1 (image, open source) Qwen-Image-2.1-Turbo (image, open source) Qwen-Image-3.0 (image) Qwen-Image-3.0-Pro (image) Qwen3-Embedding-8B / 0.6B (decision, open source) Qwen3-Reranker-4B / 0.6B (decision, open source) Qwen3.8 2.4T A95B (text, coding, open source) Qwen3.8 27B (text, coding, open source) Qwen3.8 Max (0902) (text, coding) Qwen3.8-Flash-Next (text, coding, open source) Wan 2.2 A14B (video, open source) Wan 2.5 Preview (image) Wan 2.6 Image (image) Wan 2.7 (image, video) Wan 2.7 Pro (image) Wan 3.0 (video) Wan2.6 Text to Image (image) Z-Image Base (image, open source) Z-Image Turbo (image, open source) 2 models Chinese lab known for its open-source models that are very cheap to use.
For the public: DeepSeek
Text 1 Coding 1 Image 1 Open source 2
See its models 1 model Chinese lab behind the Kimi models and assistant.
For the public: Kimi
Text 1 Coding 1 Open source 1
See its models Kimi K3 (text, coding, open source) 5 models Chinese lab that publishes the GLM models, many of them open source.
For the public: Z.ai
Text 2 Coding 2 Image 1 Video 2 Open source 4
See its models 2 models The Chinese phone maker; it publishes the MiMo models as open source.
No notable app of its own: its models are used by downloading them or via API.
Text 2 Coding 2 Open source 2
See its models 4 models Chinese lab for text, voice and video models.
No notable app of its own: its models are used by downloading them or via API.
Text 1 Coding 1 Image 1 Video 2 Open source 2
See its models 11 models TikTok's company; it publishes text, image (Seedream) and video (Seedance) models.
No notable app of its own: its models are used by downloading them or via API.
Image 9 Video 2 Open source 3
See its models 5 models Chinese short-video platform, maker of the Kling video and image generator.
For the public: Kling
Image 2 Video 3
See its models 2 models US company specializing in AI video tools.
For the public: Runway
Image 1 Video 1
See its models 4 models US video and image generation company.
For the public: Dream Machine
Image 4
See its models 17 models German lab behind the FLUX image models, several of them open.
No notable app of its own: its models are used by downloading them or via API.
Image 16 Video 1 Open source 8
See its models 3 models M MidjourneyIndependent US lab dedicated to generating artistic images.
For the public: Midjourney
Image 3
See its models 9 models I IdeogramImage generator widely used for posters and logos with text.
For the public: Ideogram
Image 9 Open source 2
See its models 9 models Maker of the Copilot assistant; publishes its own MAI models.
For the public: Copilot
Image 8 Decision 1
See its models 10 models S Stability AIMaker of Stable Diffusion, the best-known open image models.
No notable app of its own: its models are used by downloading them or via API.
Image 10 Open source 7
See its models 7 models T TencentChinese internet giant (WeChat); it publishes the Hunyuan models.
No notable app of its own: its models are used by downloading them or via API.
Image 6 Video 1 Open source 5
See its models 4 models Israeli company behind the open LTX video models.
No notable app of its own: its models are used by downloading them or via API.
Video 4 Open source 4
See its models 5 models N NVIDIAThe graphics card maker; it publishes open models.
No notable app of its own: its models are used by downloading them or via API.
Text 1 Coding 1 Image 4 Open source 5
See its models 2 models AI search engine that answers with sources; it also publishes models.
For the public: Perplexity
Decision 2 Open source 2
See its models No results.
OpenAI
Anthropic
Google
xAI
Meta
Mistral AI
Alibaba (Qwen)
DeepSeek
Moonshot AI
Z.ai (Zhipu)
Xiaomi
MiniMax
ByteDance
Kuaishou (Kling)
Runway
Luma AI
Black Forest Labs
Microsoft
Lightricks
Perplexity