Mistral AI model product
Ministral 3
A balanced model in the Ministral 3 family, Ministral 3 14B is a powerful, efficient tiny language model with vision capabilities.
Updated Aug 12, 2026. Default version: Ministral 3 (8B Reasoning 2512)
LLMBoard score29.0Ministral 3 (8B Reasoning 2512)
Coverage60%4 benchmark families
Context window262.1KTokens
Official input price$0.10Mistral API
Ministral 3 Specifications
Technical details for the model's default version.
- Version
- Ministral 3 (8B Reasoning 2512)
- Released
- Dec 4, 2025
- Knowledge cutoff
- Unknown
- Parameters
- 8B
- Context window
- 262.1K
- Max output
- 262.1K
- Inputs
- image, text
- Outputs
- text
- Open weights
- No
- License
- Apache 2.0
Ministral 3 Capability Profile
This profile uses the model's current scored version. Arena ratings and prices are shown separately.
Ministral 3 Benchmark Results
Benchmark scores for Ministral 3 (8B Reasoning 2512).
| AIME 2024 | 86.0% | 16 | 53 | 71.2% | C | |
| LiveCodeBench | 61.6% | 33 | 73 | 55.6% | C | |
| AIME 2025 | 78.7% | 72 | 114 | 37.2% | C | |
| GPQA | 66.8% | 139 | 234 | 40.8% | C |
Ministral 3 Arena Results
Preference and agent-evaluation results for the default version.
No Arena results
The default version does not have a matching Arena result yet.
Ministral 3 Runtime Performance
Provider-specific output speed and catalog latency for Ministral 3 (8B Reasoning 2512). Runtime does not affect the capability score.
| Mistral AI | 237.5 tok/s | 0.34 s | 262.1K | 262.1K |
Output Speed is generated output tokens received per second. Catalog latency is reported separately from observed provider TTFT.
Ministral 3 Pricing
Official vendor API pricing appears first, followed by individual provider offers.
- Official API
- $0.10 input, $0.10 output per 1M
- Official provider
- Mistral
- Lowest third-party
- N/A
- Tracked offerings
- 1
| Mistral | ministral-8b-latest | global | $0.10 | $0.10 | 128K |
Official prices use only the vendor's configured official Provider and positive standard USD PAYG rates. Third-party offers remain explicitly labeled.
Ministral 3 Versions
Available versions of this model. The score column identifies the version used in the overall ranking.
| Ministral 3 (14B Reasoning 2512) | N/A | 14B | 262.1K | 262.1K | No | Apache 2.0 | |
| Ministral 3 (14B Base 2512) | N/A | 14B | N/A | N/A | No | Apache 2.0 | |
| MiniStral 3 (14B Instruct 2512) | N/A | 14B | N/A | N/A | No | Apache 2.0 | |
| Ministral 3 (3B Base 2512) | N/A | 3B | N/A | N/A | No | Apache 2.0 | |
| Ministral 3 (3B Instruct 2512) | N/A | 3B | N/A | N/A | No | Apache 2.0 | |
| Ministral 3 (8B Base 2512) | N/A | 8B | N/A | N/A | No | Apache 2.0 | |
| Ministral 3 (8B Instruct 2512) | N/A | 8B | N/A | N/A | No | Apache 2.0 | |
| Ministral 3 (8B Reasoning 2512) | 29.0 | 8B | 262.1K | 262.1K | No | Apache 2.0 |
What is Ministral 3?
Key information about Ministral 3 and its available data.
A balanced model in the Ministral 3 family, Ministral 3 14B is a powerful, efficient tiny language model with vision capabilities. This model is the reasoning post-trained version, trained for reasoning tasks, making it ideal for math, coding and stem related use cases.
The Ministral 3 family is designed for edge deployment, capable of running on a wide range of hardware. Ministral 3 14B can even be deployed locally, capable of fitting in 24GB of VRAM in BF16, and less than 12GB of RAM/VRAM when quantized.
Data as of 2026-08-11.
Ministral 3 vs nearby models
Open a comparison with the three ranked models immediately above and below this model.
Models similar to Ministral 3
Recommendations prioritize the same model type and family, then the closest LLMBoard score.
MA
Mistral Small 4
Mistral AI
MA
Devstral Medium
Mistral AI
MA
Pixtral Large
Mistral AI
MA
Min istral 3
Mistral AI
MA
Magistral Small
Mistral AI
MA
Magistral Medium
Mistral AI
FAQ
Common questions about Ministral 3.
When was Ministral 3 released?
Ministral 3's default version was released on Dec 4, 2025.
How much does Ministral 3 cost?
Ministral 3's official API price is $0.10 per million input tokens and $0.10 per million output tokens via Mistral.
Who created Ministral 3?
Ministral 3 was created by Mistral AI.
What is the context window for Ministral 3?
The default version has a 262.1K token context window.
Is Ministral 3 open weight?
No. The default version is not marked as having publicly available weights.
How many API providers offer Ministral 3?
1 provider offerings are linked to the default version.
What models should I compare Ministral 3 with?
Nearby ranked alternatives include Llama 3.1 Nemotron Ultra 253B, Kimi k1.5, Gemini 2.0 Flash Thinking.