Mistral Small 4
Mistral AI / 256K
Mistral AI·256K context·Text
Hybrid model unifying Instruct, Reasoning (Magistral), and Devstral families: 40% lower completion time and 3x throughput vs Small 3.
