mistralai
Mistral: Ministral 3 3B 2512
mistralai/ministral-3b-2512
Our take
AI-assisted analysis based on the sources below. Suitability is an editorial assessment; test it on your own workload.
mistralai/ministral-3b-2512 is Mistral's smallest Ministral 3 instruct variant, presented as an efficient text-and-image-to-text model for chat and instruction following. The model card reports an FP8 checkpoint with a 3.4B language model and 0.4B vision encoder, multilingual support, system-prompt adherence, function calling and JSON output, and edge deployment fit for 8GB VRAM. OpenRouter lists text and image inputs, text output, tool and structured-output parameters, and separate benchmark indexes; treat those as measurements, not guarantees.[1][2]
Strengths
- Source-reported: compact FP8 instruct checkpoint designed for edge or local deployment; documentation says it can fit in 8GB VRAM, or less with further quantization.[2]
- Source-reported: accepts text and image input and emits text, with a documented vision encoder and multilingual coverage.[1][2]
- Source-reported: exposes function calling, JSON or structured output, system prompts and tool controls through documentation and OpenRouter parameters.[1][2]
Limitations
- Evidence is catalog and model-card only, with no independent verification in this pack; validate quality, latency and tool behavior on your workload.[1][2]
- Reported context differs: OpenRouter lists 131072 tokens while documentation describes a 256k window, so confirm the served endpoint limit.[1][2]
- Documentation recommends near-square images and low temperature for production; results may depend on settings and quantization.[2]
Where to use it
- Potentially suitable for edge or low-resource chat assistants that need concise instruction following and occasional image understanding.[1][2]
- May fit lightweight data extraction, captioning, classification, translation or short content generation, as suggested by the model card.[2]
- Could be evaluated for simple tool-using agents where structured outputs or function calling are required, keeping tool sets small per guidance.[1][2]
Sources
- [1] OpenRouter catalog · Mistral: Ministral 3 3B 2512Sources checked: Nov 13, 2026
- [2] Model documentation · Mistral: Ministral 3 3B 2512Sources checked: Nov 13, 2026
About this model
Description below is the original text supplied by OpenRouter, shown in the source language.
The smallest model in the Ministral 3 family, Ministral 3 3B is a powerful, efficient tiny language model with vision capabilities.
Specifications
- Model ID
- mistralai/ministral-3b-2512
- Creator
- mistralai
- Added to OpenRouter
- 02.12.2025
- Context length
- 131,072
- Modality
- text+image->text
- Input modalities
- text, image
- Output modalities
- text
- Token pricing
- USD per 1M tokens: in $0.100 / out $0.100
- Knowledge cutoff
- —
- Expiration date
- —
Tasks and capabilities
A grade appears only where the source supplies a matching measurement for that task, the model is active and the cohort holds at least 10 comparable models. The capability column lists facts the source declares — supported formats, reasoning and tool parameters — which are not measured quality.
| Task | Grade | Source-declared capability |
|---|---|---|
| General quality | No measurement | — |
| Coding | CC — below the B band | — |
| Agents | CC — below the B band | parameters: structured_outputs, tool_choice, tools |
| Writing | No measurement | output: text |
| Reasoning | No measurement | — |
| Document analysis | No measurement | — |
| Vision | No measurement | input: image |
| Image generation | No measurement | — |
| Audio | No measurement | — |
| Video | No measurement | — |
| Embeddings | No measurement | — |
Evaluations
| Source | Benchmark | Value | Grade | Data retrieved |
|---|---|---|---|---|
| Artificial Analysis | Coding index | 4.8 | CC — below the B band | 13.11.2026 |
| Artificial Analysis | Agentic index | 0.8 | CC — below the B band | 13.11.2026 |
| Design Arena | models · 3dwin rate 35.9% · rank 127 | 945Elo | CC — below the B band | 13.11.2026 |
| Design Arena | models · codecategorieswin rate 37.3% · rank 129 | 1,001Elo | CC — below the B band | 13.11.2026 |
| Design Arena | models · gamedevwin rate 33% · rank 136 | 939Elo | CC — below the B band | 13.11.2026 |
| Design Arena | models · websitewin rate 38.2% · rank 132 | 1,020Elo | CC — below the B band | 13.11.2026 |
How grades work
Each letter grade compares a model only against other models scored by the same source and metric in the current snapshot. S means the top 10%, A the next 20%, B the next 30%, C the rest. Models without a published measurement are shown as not rated — a supported format alone never implies measured quality.
- S — S — top 10% of the cohort
- A — A — next 20% of the cohort
- B — B — next 30% of the cohort
- C — C — below the B band
Providers
Optional live provider statistics from OpenRouter. Missing values are unknown, not zero.
| Provider | Context length | Price in /1M | Price out /1M | Latency p50 (30m, s) | Throughput p50 (30m, tok/s) | Uptime (30m) |
|---|---|---|---|---|---|---|
| Mistral | 131,072 | $0.100 | $0.100 | — | — | 100% |
| Mistral | 131,072 | $0.110 | $0.110 | — | — | 100% |
Provider data retrieved: 13.11.2026
Related articles
No published articles identify this model yet.