Mistral Small 4

mistral/mistral-small-4 · by Mistral AI · decoder-only-transformer

Pricing — 1 offering(s)

Mistral AI

About this offering: Mistral Small 4 (current Small generation). Also available as open weights; pricing here is La Plateforme API.

Input tokens

$0.10 / 1M tokens (input) Historical 2026-06-12 → 2026-09-19
$0.15 / 1M tokens (input) Current 2026-09-20 → present

Output tokens

$0.30 / 1M tokens (output) Historical 2026-06-12 → 2026-09-19
$0.60 / 1M tokens (output) Current 2026-09-20 → present

Showing the active price and any recorded history. Full pricing history is available via the paid API — see API docs.

See how Mistral Small 4 fits into a cost-aware routing setup

Capability profile

How Mistral Small 4 rates across core capability dimensions, with the task-level evidence behind each rating.

reasoning Moderate

coding Moderate

tool use Strong

Mistral Small explicitly named in function calling docs table (docs.mistral.ai/studio-api/conversations/function-calling). Model description: 'hybrid model unifying instruct, reasoning, and coding'.

instruction following Strong

context window Strong

≈128K-token context (verify for the current generation).

multilingual Strong

speed Strong

cost efficiency Strong

Operator guidance

The small, cheap Mistral tier — often available as open weights for self-hosting. Step up to Mistral Large 3 for harder reasoning/coding.

Use cases

Low-cost, high-throughput tasks; self-hosting where open weights help
Multilingual (EU) workloads on a budget

Limitations

Lower reasoning/coding ceiling than Mistral Large 3
Licence/openness varies by release — verify the exact Small 4 terms
Capability ratings are qualitative, not a single cited benchmark run