One agent session
1M input tokens and 100K output tokens, no caching.
- Input1M x $0.10 = $0.10
- Output0.1M x $0.10 = $0.01
$0.11per session
Mistral AI - Open weights
Ministral 3 3B is an open-weights model from Mistral AI that costs $0.10 per 1M input tokens and $0.10 per 1M output tokens on the first-party API, with a 256K-token context window.
Last verified from docs.mistral.ai. All sources
| Item | Price / 1M tokens |
|---|---|
| Input | $0.10 |
| Output | $0.10 |
| Cache read | $0.01 |
| Batch inputBatch requests | $0.05 |
| Batch outputBatch requests | $0.05 |
Note: Batch prices come from the pricing page's Batch mode, which applies a factor of 0.5 to standard prices. Priority Tier adds 75% and regional inference adds 10%. The docs give the context only as '256k'. No primary page states a max output. Base and Reasoning weights are also published, all Apache 2.0. The gateway id mistral/ministral-3b is unversioned and lists a 131072 context, which does not match the docs. Its price ($0.1/$0.1 with a $0.01 cache read) does match Ministral 3 3B, and the older ministral-3b-2410 was retired 2025-12-30. Re-check this mapping if the gateway changes. The model card is dated December 2, 2025; release date uses the changelog date (December 1).
ministral-3b-2512mistral/ministral-3bReasoning weights exist only as separate open weights; the API model outputs text only and is not listed in the reasoning guide. Not stated means the provider's docs do not say either way. It is not the same as No. Checked against provider docs on . Compare capabilities across models.
Worked from the standard list rates above, with the math shown so you can swap in your own token counts.
1M input tokens and 100K output tokens, no caching.
$0.11per session
10,000 requests a day at 2K input and 500 output tokens each.
$2.50per day
$75.00 per 30-day month
90% of the input is read from the prompt cache. The one-time cache write premium is not included.
$0.029per session
| Model | Input / 1M | Output / 1M |
|---|---|---|
| Ministral 3 3B (this page) | $0.10 | $0.10 |
| GLM-4-32B-0414-128KZ.ai | $0.10 | $0.10 |
| Qwen3.7-FlashAlibaba Cloud (Qwen) | $0.03 | $0.13 |
| Amazon Nova MicroAmazon | $0.035 | $0.14 |
| Command R7BCohere | $0.0375 | $0.15 |
| Ministral 3 8BMistral AI | $0.15 | $0.15 |
| Ministral 3 14BMistral AI | $0.20 | $0.20 |
| Mistral Small 4Mistral AI | $0.15 | $0.60 |
| Codestral 25.08Mistral AI | $0.30 | $0.90 |
| Mistral Large 3Mistral AI | $0.50 | $1.50 |
| Mistral Large 4Mistral AI | $1.36 | $4.18 |
Ministral 3 3B costs $0.10 per 1M input tokens and $0.10 per 1M output tokens on Mistral AI's first-party API, per docs.mistral.ai on October 17, 2026. Per 1M tokens, cache reads cost $0.01, and the batch rate is $0.05 input and $0.05 output.
Ministral 3 3B has a 256,000-token context window.
Yes. Ministral 3 3B's weights are published at https://huggingface.co/mistralai/Ministral-3-3B-Instruct-2512 under the Apache 2.0.
Mistral AI released Ministral 3 3B on December 1, 2025.
Last verified from docs.mistral.ai.
Newsletter
One email per week. New model releases, price cuts and what they mean for your bill.