Mistral Large
BalancedLast verified: June 2026· Mistral Mistral Large pricing
Flagship Mistral model. Confirm current pricing with Mistral.
- Input / 1M
- $2.00
- Output / 1M
- $6.00
- Context
- 128K tokens
- Provider
- Mistral
Pricing verified June 2026. Prices change frequently. Always confirm against the provider’s official pricing page before relying on these figures for budgeting. Official pricing →
When to choose Mistral Large
The European option, and that is usually why it is on a shortlist: EU-hosted deployment where data residency in Europe is a requirement rather than a preference. On price it sits mid-table at $2 in and $6 out, a three-times output-to-input spread that is the narrowest here alongside Llama 3.3 70B, which makes it comparatively kind to generation-heavy routes. The 128K window rules out whole-repository work. Where the EU constraint does not apply, compare its quality per dollar against the balanced rows above it and the open-weight rows below.
What Mistral Large costs in practice
A 50M/10M month is $160, just under GPT-4.1’s $180 and well above the open-weight rows. Filling the 128K window costs about 26 cents of input. The three-times output-to-input spread is the narrowest in this table alongside Llama 3.3 70B, so generation-heavy routes are treated more kindly here than almost anywhere else on the list — a genuine consideration if your workload writes more than it reads. Against that, 128K is the ceiling, and the EU-hosting question usually settles the rest.
| Scenario | Input / mo | Output / mo | Est. cost / mo |
|---|---|---|---|
| Prototype | 2M | 0.5M | $7 |
| Growing product | 50M | 10M | $160 |
| At scale | 500M | 100M | $1,600 |
Plug in your own numbers with the cost calculator.
Where Mistral Large fits in a routing setup
Routing around an EU-hosting requirement narrows the options rather than the ladder. Inside that constraint the moves are the usual ones: high-volume, low-stakes work goes to a cheaper row and this one keeps the production traffic that needs the quality. Outside it, Claude Sonnet 4.6 does the reference month for $300 against $160 here and DeepSeek-V3 for $24.50, which brackets the range this row occupies. If EU hosting turns out not to be a hard requirement, that bracket is worth revisiting.
Cheaper alternatives to Mistral Large
The models immediately below Mistral Large on blended cost, nearest first. A swap only pays if quality holds on the routes you move — see the LLM cost optimization playbook for how to prove that before you move them.
Frequently asked questions
How much does Mistral Large cost?
Mistral Large bills $2.00 per million input tokens and $6.00 per million output tokens — $3.00 per million blended at a three-to-one input:output split. A workload of 50M input and 10M output tokens per month works out at about $160. Confirm current pricing with the provider.
What is Mistral Large's context window?
Mistral Large has a 128K-token context window. One request that fills the whole window costs about $0.26 at the input rate, before the reply is billed.
Where does Mistral Large sit on price?
At $3.00 per million blended — three input tokens for every output token — Mistral Large is dearer than 8 of the 19 models tracked here and cheaper than 10. Output bills at 3x the input rate, and the nearest cheaper model on this table is Claude Haiku 4.5 at $2.00 blended.
Related concepts: context window, model routing, prompt caching, evals.