Mistral API pricing
As of November 14, 2026, the Mistral API costs $0.1 to $1.5 per million input tokens and $0.1 to $7.5 per million output tokens, across 16 Mistral models. Ministral 3 3B 2512 costs $0.1 input and $0.1 output; Ministral 3 8B 2512 costs $0.15 input and $0.15 output; Mistral Small costs $0.1 input and $0.3 output; Ministral 3 14B 2512 costs $0.2 input and $0.2 output; Mistral Tiny costs $0.25 input and $0.25 output; Open Mistral Nemo costs $0.3 input and $0.3 output; Codestral 2508 costs $0.3 input and $0.9 output; DeepSeek V4 1 Flash 1017 costs $0.3 input and $1.2 output. A typical request of 1,000 input and 500 output tokens costs $0.00015 on Ministral 3 3B 2512 and $0.00525 on Mistral Medium.
Mistral API price per 1M tokens, every model
| Model | Input | Cached input | Output | Context | Typical request |
|---|---|---|---|---|---|
Labs Leanstral 1.5.1labs-leanstral-1-5-1 | $0 | – | $0 | 262K | $0 |
Ministral 3 3B 2512ministral-3-3b-2512 | $0.1 | $0.01 | $0.1 | 131K | $0.00015 |
Ministral 3B 2512ministral-3b-2512 | $0.1 | $0.01 | $0.1 | 131K | $0.00015 |
Ministral 3 8B 2512ministral-3-8b-2512 | $0.15 | $0.015 | $0.15 | 262K | $0.000225 |
Ministral 8B 2512ministral-8b-2512 | $0.15 | $0.015 | $0.15 | 262K | $0.000225 |
Mistral Smallmistral-small | $0.1 | $0.01 | $0.3 | 32K | $0.00025 |
Ministral 3 14B 2512ministral-3-14b-2512 | $0.2 | $0.02 | $0.2 | 262K | $0.0003 |
Ministral 14B 2512ministral-14b-2512 | $0.2 | $0.02 | $0.2 | 262K | $0.0003 |
Mistral Tinymistral-tiny | $0.25 | $0.025 | $0.25 | 32K | $0.000375 |
Open Mistral Nemoopen-mistral-nemo | $0.3 | $0.03 | $0.3 | 128K | $0.00045 |
Codestral 2508codestral-2508 | $0.3 | $0.03 | $0.9 | 128K | $0.00075 |
DeepSeek V4 1 Flash 1017deepseek-v4-1-flash-1017 | $0.3 | $0.007 | $1.2 | 1.05M | $0.0009 |
Mistral Large 3mistral-large-3 | $0.5 | $0.05 | $1.5 | 262K | $0.00125 |
Mistral Large 2512mistral-large-2512 | $0.5 | $0.05 | $1.5 | 262K | $0.00125 |
Mistral Large 4mistral-large-4 | $0.7 | $0.07 | $2.1 | 1.05M | $0.00175 |
Mistral Mediummistral-medium | $1.5 | $0.15 | $7.5 | 262K | $0.00525 |
US dollars per million tokens, sorted by the cost of a typical request of 1,000 input and 500 output tokens. Also listed as mistral-large-4-0 (same as Mistral Large 4), mistral-medium-3 (same as Mistral Medium), mistral-medium-3-5 (same as Mistral Medium) and mistral-medium-3.5 (same as Mistral Medium).
How much does a Mistral API request cost?
A typical request of 1,000 input and 500 output tokens costs $0.00015 on Ministral 3 3B 2512, $0.000225 on Ministral 3 8B 2512, $0.00025 on Mistral Small, $0.0003 on Ministral 3 14B 2512, $0.000375 on Mistral Tiny, $0.00045 on Open Mistral Nemo, $0.00075 on Codestral 2508 and $0.0009 on DeepSeek V4 1 Flash 1017. A coding agent step with 50K tokens of context and 2K output tokens costs $0.0174 on DeepSeek V4 1 Flash 1017.
| Model | Short chat message 300 in, 200 out | Typical request 1K in, 500 out | Long document summary 20K in, 1K out | Coding agent step 50K in, 2K out | 1,000 typical requests |
|---|---|---|---|---|---|
| Ministral 3 3B 2512 | $0.00005 | $0.00015 | $0.0021 | $0.0052 | $0.15 |
| Ministral 3 8B 2512 | $0.000075 | $0.000225 | $0.00315 | $0.0078 | $0.23 |
| Mistral Small | $0.00009 | $0.00025 | $0.0023 | – | $0.25 |
| Ministral 3 14B 2512 | $0.0001 | $0.0003 | $0.0042 | $0.0104 | $0.30 |
| Mistral Tiny | $0.000125 | $0.000375 | $0.00525 | – | $0.38 |
| Open Mistral Nemo | $0.00015 | $0.00045 | $0.0063 | $0.0156 | $0.45 |
| Codestral 2508 | $0.00027 | $0.00075 | $0.0069 | $0.0168 | $0.75 |
| DeepSeek V4 1 Flash 1017 | $0.00033 | $0.0009 | $0.0072 | $0.0174 | $0.90 |
At standard rates, for the newest model of each Mistral family. A token is about 4 characters of English text, so 1,000 tokens is roughly 750 words. Mistral counts tokens with its own tokenizer, so the same text can be a different number of tokens.
How batch, prompt caching and long context change the price
Batch API. The price list has no batch API prices for Mistral models. See batch API pricing.
Prompt caching. All 15 priced Mistral models charge less for cached input, the part of a prompt that repeats across requests: typically 90% less. DeepSeek V4 1 Flash 1017 charges $0.007 per million cached input tokens instead of $0.3. See prompt caching pricing.
Long context. Mistral models accept 32K to 1.05M input tokens. The price list shows no long-context surcharge for Mistral: each model has one price up to its full context window. Filling a 1M-token window is still expensive, though: the input alone costs $0.30 on DeepSeek V4 1 Flash 1017.
Mistral API price changes
Since October 22, 2026, no Mistral price change has stuck. Mistral Large 4's price went up on Oct 28, 2026 (input $0.68 → $1.36 (+100%), output $2.09 → $4.18 (+100%) per 1M tokens) and went back on Nov 3, 2026, most likely a correction in the price list. 1 new model was added, most recently DeepSeek V4 1 Flash 1017 on Oct 28, 2026.
- Nov 3, 2026: Mistral Large 4 price cut: input $1.36 → $0.7 (−49%), output $4.18 → $2.1 (−50%) per 1M tokens. This undid the change of Oct 28, 2026.
- Oct 28, 2026: DeepSeek V4 1 Flash 1017 added to the price list at $0.3 input and $1.2 output per 1M tokens.
- Oct 28, 2026: Mistral Large 4 price increase: input $0.68 → $1.36 (+100%), output $2.09 → $4.18 (+100%) per 1M tokens. Reversed on Nov 3, 2026, most likely a correction.
Dated from the day a change reached the price list. Every provider: LLM API price changes.
Mistral API vs OpenAI and Claude
It depends on the model. The cheapest Mistral model, Ministral 3 3B 2512, costs $0.00015 per typical request; the cheapest OpenAI model, GPT-5 nano, costs $0.00025. DeepSeek V4 1 Flash 1017 ($0.0009) is closest in price to GPT-5.4 nano ($0.000825): about the same.
| Mistral model | Input / output | Typical request | Closest OpenAI model | Closest Claude model |
|---|---|---|---|---|
| Ministral 3 3B 2512 | $0.1 / $0.1 | $0.00015 | GPT-5 nano $0.05 / $0.4 · $0.00025 · 1.7× the price | Claude Haiku 5.5 $0.1 / $0.5 · $0.00035 · 2.3× the price |
| Ministral 3 8B 2512 | $0.15 / $0.15 | $0.000225 | GPT-5 nano $0.05 / $0.4 · $0.00025 · 1.1× the price | Claude Haiku 5.5 $0.1 / $0.5 · $0.00035 · 1.6× the price |
| Mistral Small | $0.1 / $0.3 | $0.00025 | GPT-5 nano $0.05 / $0.4 · $0.00025 · about the same | Claude Haiku 5.5 $0.1 / $0.5 · $0.00035 · 1.4× the price |
| Ministral 3 14B 2512 | $0.2 / $0.2 | $0.0003 | GPT-6 Luna $0.1 / $0.5 · $0.00035 · 1.2× the price | Claude Haiku 5.5 $0.1 / $0.5 · $0.00035 · 1.2× the price |
| Mistral Tiny | $0.25 / $0.25 | $0.000375 | GPT-6 Luna $0.1 / $0.5 · $0.00035 · about the same | Claude Haiku 5.5 $0.1 / $0.5 · $0.00035 · about the same |
| Open Mistral Nemo | $0.3 / $0.3 | $0.00045 | GPT-4o mini $0.15 / $0.6 · $0.00045 · about the same | Claude Haiku 5.5 $0.1 / $0.5 · $0.00035 · 1.3× cheaper |
| Codestral 2508 | $0.3 / $0.9 | $0.00075 | GPT-5.6 Luna $0.2 / $1.2 · $0.0008 · about the same | Claude Haiku 5.5 $0.1 / $0.5 · $0.00035 · 2.1× cheaper |
| DeepSeek V4 1 Flash 1017 | $0.3 / $1.2 | $0.0009 | GPT-5.4 nano $0.2 / $1.25 · $0.000825 · about the same | Claude Haiku 5.5 $0.1 / $0.5 · $0.00035 · 2.6× cheaper |
For each Mistral model, the model from the other provider whose typical request costs the nearest, with its input / output price per million tokens, its typical request, and how that compares with the Mistral model.
Questions and answers
How much does the Mistral API cost?
As of November 14, 2026, the Mistral API costs $0.1 to $1.5 per million input tokens and $0.1 to $7.5 per million output tokens, across 16 Mistral models. Ministral 3 3B 2512 costs $0.1 input and $0.1 output; Ministral 3 8B 2512 costs $0.15 input and $0.15 output; Mistral Small costs $0.1 input and $0.3 output; Ministral 3 14B 2512 costs $0.2 input and $0.2 output; Mistral Tiny costs $0.25 input and $0.25 output; Open Mistral Nemo costs $0.3 input and $0.3 output; Codestral 2508 costs $0.3 input and $0.9 output; DeepSeek V4 1 Flash 1017 costs $0.3 input and $1.2 output. A typical request of 1,000 input and 500 output tokens costs $0.00015 on Ministral 3 3B 2512 and $0.00525 on Mistral Medium.
Did Mistral raise its API prices?
No. Since October 22, 2026, no Mistral price change has stuck. Mistral Large 4's price went up on Oct 28, 2026 (input $0.68 → $1.36 (+100%), output $2.09 → $4.18 (+100%) per 1M tokens) and went back on Nov 3, 2026, most likely a correction in the price list. 1 new model was added, most recently DeepSeek V4 1 Flash 1017 on Oct 28, 2026.
What is the cheapest Mistral model?
Ministral 3 3B 2512 (ministral-3-3b-2512), at $0.1 per million input tokens and $0.1 per million output tokens.
How much Mistral API use does $20 buy?
It depends on how much you use it. $20 of Mistral API usage pays for about 22,222 typical requests (1,000 input and 500 output tokens) on DeepSeek V4 1 Flash 1017, or 133,333 on Ministral 3 3B 2512. At 30 typical requests a day, a month costs $0.81 on DeepSeek V4 1 Flash 1017 and $0.13 on Ministral 3 3B 2512. Long chats cost more per message than that, because every turn sends the whole conversation again as input tokens.
How much does a Mistral API request cost?
A typical request of 1,000 input and 500 output tokens costs $0.00015 on Ministral 3 3B 2512, $0.000225 on Ministral 3 8B 2512, $0.00025 on Mistral Small, $0.0003 on Ministral 3 14B 2512, $0.000375 on Mistral Tiny, $0.00045 on Open Mistral Nemo, $0.00075 on Codestral 2508 and $0.0009 on DeepSeek V4 1 Flash 1017. A coding agent step with 50K tokens of context and 2K output tokens costs $0.0174 on DeepSeek V4 1 Flash 1017.
How much does prompt caching save on the Mistral API?
All 15 priced Mistral models charge less for cached input, the part of a prompt that repeats across requests: typically 90% less. DeepSeek V4 1 Flash 1017 charges $0.007 per million cached input tokens instead of $0.3.
Does long context cost more on the Mistral API?
No. Mistral models accept 32K to 1.05M input tokens. The price list shows no long-context surcharge for Mistral: each model has one price up to its full context window. Filling a 1M-token window is still expensive, though: the input alone costs $0.30 on DeepSeek V4 1 Flash 1017.
Is the Mistral API cheaper than OpenAI?
It depends on the model. The cheapest Mistral model, Ministral 3 3B 2512, costs $0.00015 per typical request; the cheapest OpenAI model, GPT-5 nano, costs $0.00025. DeepSeek V4 1 Flash 1017 ($0.0009) is closest in price to GPT-5.4 nano ($0.000825): about the same.
Source: Mistral pricing, via the LiteLLM price list.