LLM API pricing · verified June 2026
LLM / providers / mistral

Mistral AI API Pricing

Mistral is the European option in this comparison; its publicly priced flagship, Mistral Large 3, offers caching and a batch discount at a low per-token rate with a sizeable context window.

Prices verified June 2026 · changes logged in the changelog
Heads up: The Mistral Large 3 rate reflects a promotional price rather than a settled long-term list price, so treat it as subject to change when forecasting. Mistral's wider lineup (mid, small and specialised models) does not yet have verified public per-token pricing in our data, so only the flagship is shown.
Model$ input /1M$ output /1M$ cached /1MBatch≈ $/mo *
Magistral MediumFRONTIER $2$5—−50%$550
Mistral Medium 3.5FRONTIER $1.50$7.50—−50%$525
Devstral 2FRONTIER $0.40$2—−50%$140
Mistral Large 3FRONTIER $0.50$1.50$0.05−50%$82
Magistral SmallMID $0.50$1.50—−50%$145
CodestralMID $0.30$0.90—−50%$87
Mistral Small 4MID $0.10$0.30—−50%$29
Ministral 3 (14B)BUDGET $0.20$0.20—−50%$46
Ministral 3 (8B)BUDGET $0.15$0.15—−50%$34.5
Devstral Small 2BUDGET $0.10$0.30—−50%$29
Ministral 3 (3B)BUDGET $0.10$0.10—−50%$23

* Example workload — chatbot, 100k requests/mo, 2,000 input / 300 output tokens per request, 70% of input cached. Computed by the same engine as the calculator. Batch: the −50% is Mistral's verified Batch API discount; the ≈ $/mo column is computed without it.

Prompt caching

Cached input is billed at 10% of the input rate on every Mistral model with a published cache price; Mistral Medium 3.5, Magistral Medium, Devstral 2, Mistral Small 4, Magistral Small, Codestral, Devstral Small 2, Ministral 3 (3B), Ministral 3 (8B) and Ministral 3 (14B) don't list one, so the calculator charges them the full input rate. A major lever for chatbots and agents where most of the prompt repeats — the calculator models this with your cache share.

Batch / async

The Batch API runs asynchronous jobs at a verified −50% on both input and output across all models we track — flip the Batch toggle in the calculator to model it.

Context window

Mistral Large 3, Mistral Medium 3.5, Devstral 2, Mistral Small 4, Devstral Small 2, Ministral 3 (3B), Ministral 3 (8B) and Ministral 3 (14B) run a verified 256k-token context window; Magistral Medium, Magistral Small and Codestral are 128k tokens.

When Mistral is worth it

Use caseVerdict
You want an EU-based vendor for data-residency reasonsMistral is the European option here
A single flagship model covers your needsMistral Large 3 with caching and batch fits
You need a budget or specialised tier with a verified public priceOnly the flagship is currently verifiable
Is Mistral the right price for your workload?
The calculator puts these eleven models next to the other 17 we track — at your volume, token mix and cache share.
Open calculator

Frequently asked questions

Only the flagship tier has a verified public per-token price in our data. Mistral's wider lineup (mid, small and specialised models) is expected to be added once its public per-token rates and context windows are confirmed.
It reflects a promotional rate rather than a settled long-term list price, so treat it as subject to change for long-running budgets. The current figure and whether caching and batch apply are shown in the table above.
All 28 models → OpenAI pricing → Anthropic pricing → Gemini pricing → DeepSeek pricing → Grok pricing → Mistral alternatives → Cheapest LLM API → Price changelog →