mistralai
Mistral: Mistral Medium 3.5
mistralai/mistral-medium-3-5
Our take
AI-assisted analysis based on the sources below. Suitability is an editorial assessment; test it on your own workload.
mistralai/mistral-medium-3-5 is Mistral AI's dense 128B unified model that, per its model card, handles instruction following, optional reasoning, and coding in one set of weights. It accepts text, image, and file inputs (file per the OpenRouter catalog; the card lists text and image) and outputs text over a 256k context window. Reasoning effort is configurable per request. Mistral positions it for agentic workflows and says it replaces Medium 3.1, Magistral, and Devstral 2 in its own products. Vendor-reported benchmarks are unverified; validate on your workload.[1][2]
Strengths
- Single dense 128B checkpoint covering instruct, reasoning, and coding, with reasoning effort settable per request: high for complex or agentic tasks, none for fast replies.[1][2]
- Multimodal input (text, image, and per the OpenRouter listing files) over a 256k context window; the card notes a vision encoder trained from scratch for variable image sizes and aspect ratios.[1][2]
- Mistral reports native function calling, JSON and structured output support, strong system-prompt adherence, and coverage of dozens of languages; tool and structured-output parameters are exposed in the catalog.[1][2]
Limitations
- Evidence is a vendor model card plus an aggregator catalog; benchmark figures are self-reported and not independently verified here, so confirm quality and cost on your own workload.[1][2]
- The card warns that an earlier incorrect Transformers configuration degraded long-context performance and that GGUFs built from it remain affected; deployments need the fixed config.[2]
- Released under a modified MIT license with exceptions for high-revenue companies, so commercial use may need legal review; output is text-only despite multimodal input.[2]
Where to use it
- Agentic coding and tool-calling pipelines: Mistral designed the model for this and ships it in its Vibe coding agent, making it a plausible candidate for similar workflows.[1][2]
- Long-context analysis combining documents, images, and instructions, where one model covering chat, reasoning, and vision simplifies the stack; worth piloting on representative tasks.[1][2]
- Multilingual assistants across the many languages the card lists, including Spanish and Portuguese, subject to per-language evaluation.[2]
Sources
- [1] OpenRouter catalog · Mistral: Mistral Medium 3.5Sources checked: Nov 13, 2026
- [2] Model documentation · Mistral: Mistral Medium 3.5Sources checked: Nov 13, 2026
About this model
Description below is the original text supplied by OpenRouter, shown in the source language.
Mistral Medium 3.5 is a dense 128B instruction-following model from Mistral AI. It supports text and image inputs with text output, and is designed for agentic workflows, coding, and complex...
Specifications
- Model ID
- mistralai/mistral-medium-3-5
- Canonical ID
- mistralai/mistral-medium-3.5-20260430
- Creator
- mistralai
- Added to OpenRouter
- 30.04.2026
- Context length
- 262,144
- Modality
- text+image+file->text
- Input modalities
- text, image, file
- Output modalities
- text
- Token pricing
- USD per 1M tokens: in $1.5 / out $7.5
- 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 | CC — below the B band | reasoning: optional · default high · efforts: high, none · parameters: include_reasoning, reasoning, reasoning_effort |
| Coding | BB — next 30% of the cohort | — |
| Agents | CC — below the B band | parameters: structured_outputs, tool_choice, tools |
| Writing | No measurement | output: text |
| Reasoning | No measurement | reasoning: optional · default high · efforts: high, none · parameters: include_reasoning, reasoning, reasoning_effort |
| Document analysis | No measurement | input: file |
| 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 | Intelligence index | 4.6 | CC — below the B band | 13.11.2026 |
| Artificial Analysis | Coding index | 46.9 | BB — next 30% of the cohort | 13.11.2026 |
| Artificial Analysis | Agentic index | 7 | 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 | $1.5 | $7.5 | — | — | 100% |
| Mistral | 262,144 | $1.65 | $8.25 | — | — | 100% |
Provider data retrieved: 13.11.2026
Related articles
No published articles identify this model yet.