AI model directory
A neutral reference table of widely deployed AI models: who makes them, what they are good at, and how they are licensed.
Last updated Aug 22, 2026
| Models | Developer | Modality | Access | Best for |
|---|---|---|---|---|
| Claude | Anthropic | Text, images, documents | Proprietary API | Long-document analysis, coding agents and tool use where instruction-following matters more than raw speed. |
| GPT | OpenAI | Text, images, audio | Proprietary API | General-purpose assistants and the widest third-party ecosystem — most integrations target it first. |
| Gemini | Google DeepMind | Text, images, audio, video | Proprietary API | Multimodal input — video and audio understanding — plus tight integration with Google Cloud and Workspace. |
| Llama | Meta | Text, images | Open weights | Self-hosting and fine-tuning. Note the licence is a custom community licence, not OSI open source. |
| Qwen | Alibaba Cloud | Text, images | Open weights | Open-weight deployments that need strong Chinese and multilingual coverage, in many parameter sizes. |
| DeepSeek | DeepSeek | Text | Open weights | Reasoning and code at low cost per token; the weights are downloadable under a permissive licence. |
| Mistral | Mistral AI | Text, images | Mixed | European data residency, plus a mix of Apache-2.0 open models and larger commercial ones. |
| Grok | xAI | Text, images | Proprietary API | Products that want live X/Twitter context; older versions have been released as open weights. |
Specifications change often. Each row links to the vendor's own documentation, which is the authoritative source.