Every AI model. One place.

Meta's latest compact model. Great for general chat and coding.

Meta's flagship open-source model. Excellent reasoning.

Meta's largest open model. State-of-the-art performance.

Mistral AI's efficient 7B model. Fast and capable.

Mixture-of-experts model. High quality at efficient inference.

Microsoft's compact model. Punches above its weight.

Google's open model. Strong multilingual and reasoning.

DeepSeek's reasoning model. Excellent for math and logic.

Alibaba's multilingual model. Strong on Chinese and English.

Specialized coding model. Supports 300+ languages.

Meta's code-specialized Llama. Great for code completion.

Qwen's code-tuned variant. Excellent code generation.

Generate images locally. No API key needed.

Vision-language model. Understands images and text.

OpenAI's speech recognition. Transcribe audio locally.

High-quality text embeddings for RAG and search.

OpenAI's flagship multimodal model.

Fast and affordable. Great for most tasks.

OpenAI's reasoning model for complex problems.

OpenAI's image generation model.

Anthropic's balanced model. Great reasoning.

Anthropic's most capable model.

Fast and affordable. Excellent for quick tasks.

Google's fast multimodal model.

Google's most capable. 1M token context.

Google's image generation model.

Mistral's most capable cloud model.

DeepSeek's flagship cloud model.

Cloud-hosted reasoning model.