@MistralDevs
š„ Meet Mistral Small 4: One model to do it all. ā” 128 experts, 119B total parameters, 256k context window ā” Configurable Reasoning ā” Apache 2.0 ā” 40% faster, 3x more throughput Our first model to unify the capabilities of our flagship models into a single, versatile model. https://t.co/2M1VNaDkRz