Mistral API pricing

Mistral runs a compact, clearly-tiered lineup — Mistral Small in the Budget tier and Mistral Medium in the Standard tier — priced competitively against the larger US providers for comparable capability. See the live table above for current rates.

Small vs. Medium

Mistral Small is a genuinely capable Budget-tier model, well suited to classification, extraction, summarization, and moderate chat workloads without the compromises sometimes seen in the cheapest tier from other providers. Mistral Medium steps up to Standard-tier general-purpose capability at a price that typically undercuts the equivalent Standard-tier models from OpenAI and Google — worth checking directly before defaulting to a larger provider out of habit.

Why teams choose Mistral specifically

Beyond raw price, Mistral is frequently chosen for two reasons that don't show up on a pricing table:

• European data residency and regulatory alignment. As a France-based provider, Mistral is a common default for teams with EU data-handling requirements or a preference for a non-US model provider — confirm current data-residency terms directly for your specific compliance needs.
• Efficient small-model performance. Mistral has consistently punched above its parameter count in independent evaluations, which is part of why Mistral Small competes credibly with Budget-tier offerings from providers with larger overall model families.

Practical guidance

Because Mistral's tiers are simpler than a full Budget/Standard/Premium spread, the routing decision is largely binary: use Small for high-volume, lower-complexity work, and move to Medium when task complexity genuinely requires it. Check current standing on the LiveBench cost-vs-quality chart before assuming Medium is necessary — Small frequently sits closer to the value frontier than its Budget-tier label might suggest.

Rate limits

Mistral runs a five-tier structure from a free tier up through a top tier unlocked by meaningful cumulative spend. One structural quirk worth knowing: Mistral enforces limits per requests-per-second, not per-minute like most peers, alongside separate tokens-per-minute and monthly token caps — a workload that bursts briefly can hit the per-second cap even well under a per-minute budget. Exact current figures require logging into Mistral's admin console for your account.

See how this compares to the other 8 providers we track on the rate limits page.

When Mistral vs. a competitor

For general-purpose Standard-tier work, compare Mistral Medium directly against OpenAI and Google Standard-tier models on the compare page — price and capability are close enough that the comparison is worth making explicitly rather than assuming a larger provider is the safer default. For pure lowest-cost routing, also check Mistral Small against DeepSeek and Qwen.

See current Small-vs-Medium pricing live above, or put Mistral head-to-head with any competitor on compare.