Mistral AI · Flagship

Mistral Medium 3.5 API pricing and specs

Mistral's main model for agents and coding, released as open weights, which replaces Medium 3.1, the Magistral models and Devstral 2.

Input

⁨$1.50⁩

Cached input

—

Output

⁨$7.50⁩

US dollars per million tokens, standard tier, short-prompt rate

Mistral publishes no per-model cached-input price. Enterprise API access costs 75% more than list.

Spec sheet

API model string: mistral-medium-3-5
Provider: Mistral AI
Context: 256K tokens
Max output: —
Released: April 28, 2026
Batch API: 50% off, results within hours
Documented agent features: Tool callingStructured outputAdjustable reasoningImage input

What a task costs

Four reference workloads priced at this model's list rates, with caching applied after the first call. Your numbers will differ; the calculator lets you change every assumption.

Workload	Per task	Per 1,000 tasks
Support bot 4 calls, short tool lookups	⁨$0.068⁩	⁨$68⁩
Coding agent 30 calls, files and test output	⁨$5.31⁩	⁨$5,310⁩
Browser agent 25 calls, screenshots and page text	⁨$2.18⁩	⁨$2,175⁩
Research agent 15 calls, long search results	⁨$1.65⁩	⁨$1,654⁩