Mistral Medium 3.5
A solid pick for self-hosting and full data control.
Overview
Mistral Medium 3.5, released on 29 April 2026, is Mistral's first merged model: a single set of weights covers instruction following, reasoning and coding, where you previously had to choose between Medium, Magistral and Devstral. It replaces Medium 3.1 and Magistral in Le Chat, and Devstral 2 in the Vibe coding agent. Reasoning effort is set per request, from quick replies to long agentic tasks. It is a dense 128-billion-parameter model with a vision encoder and a 256k-token window. It is served on Mistral's API and released as open weights on Hugging Face under a modified MIT license: commercial use is allowed, with an exception for very large-revenue companies, worth reading before a large-enterprise rollout. For a European IT leader, it is the reference Mistral model when quality comes first: self-hostable on a multi-GPU server, or available through a French vendor's API. For high-volume, less demanding work, Mistral Small 4 costs ten times less.
Skill profile
Not disclosed
Three independent measures
What people prefer, what the model can do, what the market uses. Each source has its biases and none is enough on its own: we show them side by side, never blended.
Strengths
- World-class on Arena
- Has a free tier
- Open-source and self-hostable
Who is it for
- you want to control cost or self-host
- your data must stay in the EU
Ideal use cases
- High-end sovereign enterprise assistant
- Coding agents and repository review
- Long-document analysis with images
Access & availability
Key specifications
What it costs per month
Estimate from sourced API prices and a hardware tier for self-hosting. Assumptions are editable.
Assumptions: 36-month amortisation, 0.25 €/kWh, 22 days/month, 1 $ = 0.92 €. Excludes labour, software licences, redundancy.
nAIvigate Trust Index
What the vendor offers and documents for professional use in Europe. An undocumented criterion counts as zero and stays visible: the score rises when the vendor publishes, not when we assume.
19/19 criteria documented · 100 % · 12 of them set by rule
See the criteria▾
Trust Index badge
Show this score in your docs, README or security page. It updates automatically.
[](https://naivigate.com/en/model/mistral-medium-3-5)Data & compliance
What we know, line by line, with source and date. These lines feed the Data and Enterprise families of the Trust Index above. GDPR compliance depends on your processing, not on the model alone.
| EU data residency | Yes |
| Data processing agreement (DPA) | N/A· self-hosted: data never leaves |
| Transfer outside the EU | No |
| Training on your data | No |
| Configurable retention | N/A· self-hosted: data never leaves |
| Zero Data Retention | N/A· self-hosted: data never leaves |
| Self-hostable | Yes |
| Declared hosting | EU |
| Anonymisable data | Yes |
Quality of this card’s data
This score measures the quality of the data on this card: price, context, Elo, licence, documented, sourced and fresh. It measures neither the model nor its compliance: for that, see the Trust Index above.
7/8 documented · 4/8 primary sources
Run it locally
| Quantization | Disk | RAM / VRAM | Typical hardware |
|---|---|---|---|
| Q4 · recommended | 73 GB | 82 GB | 96-128GB Mac Studio / 4× GPUs |
| Q8 · balanced | 137 GB | 153 GB | Server GPU infra (H100/A100…) |
| FP16 · max quality | 256 GB | 284 GB | Server GPU infra (H100/A100…) |
Estimates for a moderate context. Long contexts need more RAM (KV cache).
Deploy
Copy-ready commands generated from this card. Adjust context length and GPU count to your hardware.
OpenAI-compatible server for production on NVIDIA GPUs.
pip install vllm
vllm serve mistralai/Mistral-Medium-3.5-128B \
--max-model-len 32768 \
--tensor-parallel-size 2 \
--dtype autoAdvanced data · for expertsArchitecture, modalities, detailed cost, full benchmarks▾
| Arena Elo | 1426 |
| MMLU | — |
| GPQA | — |
| HumanEval | — |
| SWE-Bench | — |
| MATH | — |