invokeable()
← comparisons / mistral-medium-2505-vs-mistral-medium-2604

Mistral Medium 3 vs Mistral Medium 3.5

Side-by-side from the Invokeable directory — provider-attributed pricing, canonical IDs, draft data flagged. Open as a shareable card →

01Head to head

Mistral Medium 3Mistral Medium 3.5
CreatorMistral AIMistral AI
Released2025-05-072026-04-29
Context131K262K
Max output131K262K
Input $/M——
Output $/M——
Knowledge cutoff2025-05—
Modalitiestext + image → texttext + image → text
Capabilitiestool-usetool-use, extended-thinking, structured-output
ProvidersMistral AI (La Plateforme)Mistral AI (La Plateforme)

02Benchmarks · vendor-reported unless verified

Mistral Medium 3Mistral Medium 3.5
Artificial Analysis Coding Index (index)13.6—
SWE-bench Verified (resolved)—77.6
SciCode33.1%—
Terminal-Bench Hard (success rate)3.8—
τ³-Telecom (accuracy)—91.4