Mistral Small 4 merges reasoning, vision and coding into one Apache 2.0 model
Mistral Small 4 is a 119B-parameter mixture-of-experts model with about 6B active parameters per token, a 256k context window and image input, released under Apache 2.0. It merges Mistral's separate reasoning, vision and coding/agent lines into one model with configurable reasoning effort. For teams wanting a permissive European open model, it simplifies deployment by replacing several specialist checkpoints with one.
Read the full story at Mistral AI ↗Like, save or discuss on Buzz
