Mistral Medium 3.5 lands — capstone on a six-week release blitz
Mistral Medium 3.5 (Apr 29) is a frontier multimodal model targeted at agentic and coding workloads. It's the headline at the end of a stretch where Mistral shipped Small 4 (unifying Magistral/Pixtral/Devstral), Voxtral TTS, Leanstral for formal proofs, and the Forge enterprise platform — all between March 16 and end of April.
Mistral Medium 3.5 is positioned as a frontier multimodal model optimized for agentic and coding use cases. The interesting structural move from this release cycle is Mistral Small 4 (March 16), which unifies what were three separate model lines — Magistral (reasoning), Pixtral (multimodal vision), and Devstral (agentic coding) — into one versatile system. The model proliferation that defined early 2026 is starting to collapse back into consolidated checkpoints.
Also notable from this window: Voxtral TTS (March 23) — Mistral's first open-weights audio model, nine languages, positioned as a direct shot at ElevenLabs. Leanstral (March 16) — the first open-source agent for Lean 4 formal proof engineering; the pitch is generating code plus a machine-checkable proof of correctness rather than just code-and-pray. Mistral Forge (March 17 at NVIDIA GTC) — enterprise platform for training frontier-grade models on proprietary data.
Mistral Large 3 (December 2025) — 41B active / 675B total MoE under Apache 2.0 — remains the flagship.