Ministral 3 8B
Ministral 3 8B is the middle size in the Ministral 3 range. Its language component has 8.4B parameters, plus a 0.4B vision encoder. It can process image and text requests, return JSON and call defined tools. Mistral offers both an API model and Apache 2.0 weights.
Choose the checkpoint and precision
Mistral supplies an FP8 instruction checkpoint and reports that it fits in 12GB of VRAM. Separate Base and Reasoning checkpoints serve different uses. The family card lists BF16 for Base and Reasoning, so memory figures for FP8 Instruct should not be copied to those releases.Ministral 3 8B Instruct model card
Use the instruction API
The hosted model identifier is ministral-8b-2512. Its model page lists function calling, structured output, document questions, prefix completion and batch processing. The Base and Reasoning names here identify published weight variants; they are not additional confirmed serverless aliases.Ministral 3 8B API model page
Check the device workload
Practical guidance. For a local document assistant, compare 8B against 3B on difficult layouts and against 14B on the same hardware budget. Include long prompts and more than one active request. This reveals whether the middle size provides enough extra accuracy without losing the response time the application needs.
Specifications & access
| Detail | Value and conditions |
|---|---|
| Access | API, Downloadable |
| Model identifiers | ministral-8b-2512Ministral-3-8B-Instruct-2512Ministral-3-8B-Base-2512Ministral-3-8B-Reasoning-2512 |
| Context window | 262,144 tokens The official Instruct serving example sets a maximum model length of 262144; available capacity depends on deployment memory. Ministral 3 8B Instruct model card |
| Release | 2 December 2025; general availabilityMinistral 3 8B API model page |
| Components | 8.4B language model and 0.4B vision encoderMinistral 3 8B Instruct model card |
| License | Apache 2.0Ministral 3 8B API model page |
Related models
- Mistral Small 4
Mistral AI · Mistral Small
An Apache-licensed multimodal model that combines instruction following, reasoning and coding with selectable reasoning modes.
Launch date not confirmedAPI · Downloadable - Ministral 3 3B
Mistral AI · Ministral 3
The smallest Ministral 3 release pairs a 3.4B language model with a 0.4B vision encoder and offers multiple training variants.
Launch date not confirmedAPI · Downloadable - Ministral 3 14B
Mistral AI · Ministral 3
The largest Ministral 3 model provides 14B text-and-image checkpoints for base training, instruction following and reasoning.
Launch date not confirmedAPI · Downloadable