mistralai
Mistral: Ministral 3 14B 2512
mistralai/ministral-14b-2512
Our take
AI-assisted analysis based on the sources below. Suitability is an editorial assessment; test it on your own workload.
Ministral 3 14B Instruct 2512 is the largest entry in Mistral's Ministral 3 family: an FP8 instruction-tuned release pairing a 13.5B language model with a 0.4B vision encoder, so it accepts text and images and returns text. The model card advertises a 256k context window, multilingual support, native function calling with JSON output, and an Apache 2.0 license, and states the FP8 weights can fit in about 24 GB of VRAM for local deployment. Vendor-published benchmarks compare it favorably with similar-sized models, but those are publisher-reported results that still require validation on real workloads.[1][2]
Strengths
- Multimodal input: the card and the OpenRouter listing describe text and image input with text output, plus a 256k context window for long documents.[1][2]
- Agentic features: the card claims native function calling and JSON output, and the catalog exposes tools, tool_choice and structured_outputs parameters.[1][2]
- Deployment flexibility: the card says FP8 weights fit in roughly 24 GB of VRAM (less if further quantized) and the Apache 2.0 license permits commercial use and modification.[2]
Limitations
- Evidence is limited to the vendor model card and an OpenRouter catalog entry; no independent evaluations were included, and claims like 'best-in-class' or comparability to Mistral Small 3.2 24B are unverified vendor statements.[1][2]
- This is the Instruct FP8 variant: the separate Reasoning variant's behavior cannot be assumed, and the OpenRouter entry lists reasoning as null.[1][2]
- Knowledge cutoff is not specified, and the card's recommendations (temperature below 0.1 for production, near-square image aspect ratios) suggest tuning is needed; validate on your own workload.[1][2]
Where to use it
- Private or on-premises chat assistants in environments with limited GPU capacity, a deployment scenario the vendor explicitly targets.[2]
- Local agentic workflows that rely on tool and function calling, which the card highlights as a core capability.[2]
- Image-aware tasks such as visual question answering, plus fine-tuning and specialization under the permissive license.[2]
Sources
- [1] OpenRouter catalog · Mistral: Ministral 3 14B 2512Sources checked: Nov 13, 2026
- [2] Model documentation · Mistral: Ministral 3 14B 2512Sources checked: Nov 13, 2026
About this model
Description below is the original text supplied by OpenRouter, shown in the source language.
The largest model in the Ministral 3 family, Ministral 3 14B offers frontier capabilities and performance comparable to its larger Mistral Small 3.2 24B counterpart. A powerful and efficient language...
Specifications
- Model ID
- mistralai/ministral-14b-2512
- Creator
- mistralai
- Added to OpenRouter
- 02.12.2025
- Context length
- 262,144
- Modality
- text+image->text
- Input modalities
- text, image
- Output modalities
- text
- Token pricing
- USD per 1M tokens: in $0.200 / out $0.200
- 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 | 14.4 | CC — below the B band | 13.11.2026 |
| Artificial Analysis | Agentic index | 1.1 | CC — below the B band | 13.11.2026 |
| Design Arena | models · 3dwin rate 39.6% · rank 122 | 973Elo | CC — below the B band | 13.11.2026 |
| Design Arena | models · codecategorieswin rate 44% · rank 120 | 1,056Elo | CC — below the B band | 13.11.2026 |
| Design Arena | models · gamedevwin rate 43.6% · rank 121 | 1,028Elo | CC — below the B band | 13.11.2026 |
| Design Arena | models · websitewin rate 44.8% · rank 119 | 1,074Elo | 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 | 262,144 | $0.200 | $0.200 | — | — | 100% |
| Mistral | 262,144 | $0.220 | $0.220 | — | — | 100% |
Provider data retrieved: 13.11.2026
Related articles
No published articles identify this model yet.