Provider pricing
Mistral AI API pricing
Mistral has 62 priced models in the public snapshot. For 1,000 input and 500 output tokens, Mistral NeMo has the lowest recorded request cost at $0.000035, with $0.02 input and $0.03 output per million tokens.
Best for
European teams, multilingual applications, or those needing EU data residency.
Strengths
- European provider — GDPR-compliant, EU data residency
- Strong multilingual performance across European languages
- Mistral Small at $0.10/M — among the cheapest options
- Open-weight models available for self-hosting
Tracked Mistral models
| Model | Input $/M | Output $/M | Context | Tier | Best for |
|---|---|---|---|---|---|
| Mistral Large 4 | $0.680 | $2.09 | 1,048,576 | Mid-tier | Mistral flagship with 1M context |
| Mistral Large 3 | $0.500 | $2.00 | 131,000 | Flagship | European provider, multilingual |
| Mistral Medium 3 | $0.400 | $2.00 | 131,000 | Mid-tier | Balanced workloads, function calling |
| Mistral Small 3.1 | $0.150 | $0.600 | 128,000 | Budget | High-throughput budget tasks |
| Magistral Medium 2 | $2.00 | $5.00 | 128,000 | Premium | Frontier European reasoning model with extended chain-of-thought |
| Magistral Small | $0.500 | $1.50 | 128,000 | Mid-tier | Open-weight 24B reasoning model — math, science, code |
| Codestral 25.10 | $0.300 | $0.900 | 256,000 | Mid-tier | Code generation, completion, and FIM across 80+ languages |
| Codestral Embed | $0.150 | $0.150 | 8,192 | Budget | Code embeddings for retrieval and semantic search |
| Devstral Medium | $0.400 | $2.00 | 256,000 | Mid-tier | Agentic coding — long-context repo edits and tool use |
| Devstral Small | $0.100 | $0.300 | 128,000 | Budget | Open-weight agentic coder for IDEs and local dev |
| Ministral 8B | $0.100 | $0.100 | 128,000 | Budget | Edge-class 8B model — low-latency on-device chat |
| Ministral 3B | $0.040 | $0.040 | 128,000 | Budget | Ultra-cheap 3B model for classification and routing |
| Pixtral Large | $2.00 | $6.00 | 128,000 | Flagship | Multimodal flagship — vision, charts, document understanding |
| Pixtral 12B | $0.150 | $0.150 | 128,000 | Open Source | Open-weight 12B vision-language model |
| Mistral OCR | $1.00 | $1.00 | 0 | Mid-tier | Document OCR with layout and table extraction |
| Mistral Embed 2 | $0.100 | $0.100 | 32,000 | Budget | Multilingual embeddings for RAG and clustering |
| Mistral Moderation | $0.100 | $0.100 | 32,000 | Budget | Content moderation across 9 safety categories |
| Mistral Saba | $0.200 | $0.600 | 32,000 | Mid-tier | Arabic + South-Asian language specialist |
| Mistral NeMo | $0.150 | $0.150 | 128,000 | Open Source | Open-weight 12B model co-developed with NVIDIA |
| Mixtral 8x22B | $2.00 | $6.00 | 64,000 | Open Source | Sparse mixture-of-experts open-weight model |
| Mistral Large 2 | $2.00 | $6.00 | 128,000 | Flagship | Previous-gen Mistral flagship — strong reasoning and multilingual |
| Mistral 7B Instruct | $0.070 | $0.070 | 32,000 | Open Source | Classic open-weight 7B instruct model |
| Mixtral 8x7B | $0.240 | $0.240 | 32,000 | Open Source | Sparse MoE 8x7B — efficient open-weight workhorse |
| Mistral Large 4 | $0.600 | $2.40 | 256,000 | Flagship | Next-gen European flagship — multilingual and function calling |
| Codestral 26 | $0.350 | $1.00 | 256,000 | Mid-tier | Refreshed Codestral for completion, FIM, and 80+ languages |
Last updated: · Source: Mistral pricing · Methodology
Frequently asked questions
How much does the Mistral API cost?
Mistral has 62 priced models in the public snapshot. For 1,000 input and 500 output tokens, Mistral NeMo has the lowest recorded request cost at $0.000035, with $0.02 input and $0.03 output per million tokens.
Which Mistral model is cheapest?
Mistral NeMo is the lowest-cost recorded model for the 1,000-input/500-output example. Prices are $0.02 input and $0.03 output per million tokens; different workloads may change the ranking.
Does Mistral have cached or batch prices?
Cached input: not separately recorded. Batch: not recorded in this dataset. Eligibility and cache-write charges depend on the provider; these discounts are not automatically stacked.
How often are provider prices verified?
The live pricing table synchronizes stored model records daily; public file snapshots update when the site is rebuilt. Synchronization is not independent provider verification. Snapshot updated 2026-10-31; only explicit per-model dates are verification records.