Ministral 3 3B
Ministral 3 3B brings text and image input to the smallest model in the Ministral 3 family. The name refers to a 3.4B language model paired with a 0.4B vision encoder. It has API access and Apache 2.0 weights for local extraction, captions and short responses.
Choose the checkpoint and precision
The FP8 Instruct checkpoint is documented as fitting in 8GB of VRAM. Mistral also publishes Base and Reasoning versions. The card recommends a small, clearly defined tool set. This is a useful constraint for a model that must choose between actions on a limited device.Ministral 3 3B Instruct model card
Use the instruction API
The hosted model identifier is ministral-3b-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 3B API model page
Check the device workload
Practical guidance. Start with narrow requests, such as extracting a few fields or describing one image. Define the output format and measure missed fields or incorrect tool choices. Escalation to a larger model should follow those observed errors. A small language component does not make every long-context or vision request inexpensive.
Specifications & access
| Detail | Value and conditions |
|---|---|
| Access | API, Downloadable |
| Model identifiers | ministral-3b-2512Ministral-3-3B-Instruct-2512Ministral-3-3B-Base-2512Ministral-3-3B-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 3B Instruct model card |
| Release | 2 December 2025; general availabilityMinistral 3 3B API model page |
| Components | 3.4B language model and 0.4B vision encoderMinistral 3 3B Instruct model card |
| License | Apache 2.0Ministral 3 3B 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 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 - Ministral 3 8B
Mistral AI · Ministral 3
An 8B multimodal Ministral release with separate pretrained, instruction and reasoning checkpoints for local deployment.
Launch date not confirmedAPI · Downloadable