Skip to content

Mistral Large 4 1T A52B

mistralai/Mistral-Large-4-1T-A52B
Open Source · chat · open-weights
Open Alert me on changes
Context
1M
Max output
256K
Weights
Open
API $/1M
$0.68 / $2.09
Modalities
text · image
Released
30 Sept 2026
Download image Share on X Share on LinkedIn
AI summary
● machine-written

Mistral Large 4: 1.05T mixture-of-experts model with 1M context window

Mistral Large 4 is a multimodal mixture-of-experts language model released in October 2026 with 1.05 trillion total parameters and approximately 49-52 billion active parameters. It supports text and image inputs, offers a 1M-token context window, and is available through Mistral AI's API. The model is positioned for reasoning, coding, and agentic workloads.

What's new
  • 1M-token context window with up to 256K output tokens
  • Multimodal support for text and image inputs
  • Mixture-of-experts architecture with 49-52B active parameters
  • Tool calling and structured outputs support
  • Open-weights release available on Hugging Face
Best for
Reasoning and complex problem-solving tasksCode generation and software developmentAgentic workloads requiring tool callingMultimodal tasks combining text and image understanding
Sources

Source: https://huggingface.co/mistralai/Mistral-Large-4-1T-A52B