Mistral AI is a French AI company headquartered in Paris. As an EU-native company, Mistral offers enterprises in the DACH region with strict privacy and compliance requirements models that can be operated in the EU – via its own cloud in Paris, via Microsoft Foundry (Data Zone Standard EU) or by self-hosting open weights. The portfolio ranges from language models (Medium 3.5, Small 4, Large 3) through audio (Voxtral) and code (Codestral, Devstral) to document AI (OCR 4.1).
The current reference models are Mistral Medium 3.5 (flagship for agents, coding and reasoning), Mistral Small 4 (efficient open-source model under Apache 2.0), Mistral Large 3 (open-weight model for demanding tasks) and Mistral OCR 4.1 for documents. As an AI consultancy from Rosenheim we support model selection, GDPR assessment and operations, including through CompanyGPT. This overview reflects the state as of October 2026.
Current models at a glance
| Model | Release | Context | Licence | Focus |
|---|---|---|---|---|
| Mistral Medium 3.5 | 29 April 2026 | 256k | Open weights (modified MIT licence) | Agents, coding, reasoning |
| Mistral Small 4 | 16 March 2026 | 256k | Apache 2.0 | Efficient standard workloads, self-hosting |
| Mistral Large 3 | December 2025 | – | Mistral Research License | Complex enterprise applications, agentic workflows |
| Mistral OCR 4.1 | GA 31 August 2026 | – | API | Document processing |
| Leanstral 1.5 | 2 July 2026 | – | Apache 2.0 | Formal proofs in Lean 4 (self-hosting only) |
Mistral Medium 3.5
With Medium 3.5 (29 April 2026) Mistral released its first merged flagship model. It combines the previously separate models Magistral (reasoning) and Devstral 2 (agentic coding) in one dense 128B model with 256k context, a switchable reasoning mode and native image processing. According to Mistral it reaches 77.6% on SWE-Bench Verified. The weights are available under a modified MIT licence. Medium 3.5 also powers the Vibe agent (see below). EU data residency applies by default; since 28 May 2026 the model is also available in Microsoft Copilot Studio with in-region processing for EU customers.
Mistral Small 4
Small 4 (16 March 2026) uses a sparse MoE architecture with 119 billion parameters, of which only about 6 billion are active (4 experts). It offers 256k context, native vision and the reasoning_effort parameter to control reasoning depth at runtime. According to Mistral, latency drops by about 40% compared with Small 3 while throughput triples. The Apache 2.0 licence allows unrestricted self-hosting.
Mistral Large 3 and specialised models
Mistral Large 3 has 675 billion total parameters (41 billion active) and is designed for reasoning, multimodal processing and agentic workflows. Specialised models include:
- Codestral: code generation and analysis
- Voxtral: audio and speech-to-text for multilingual transcription
- Magistral Medium 1.2: reasoning (superseded by Medium 3.5)
- Ministral 3 family: edge models in 14B, 8B and 3B for on-device use
Mistral OCR 4.1
On 31 August 2026 Mistral declared OCR 4.1 (mistral-ocr-4-1) generally available; the first release dates from 16 July 2026 per the changelog. The aliases mistral-ocr-latest and mistral-ocr-4 point to OCR 4.1 – anyone using them is already on the new model and should re-check their pipeline outputs. New per Mistral’s documentation:
- Paragraph-level bounding boxes – native, without a downstream layout analysis.
- Structural block labels that type layout elements.
- Block-level confidence scores: the
confidence_scores_granularityparameter supportspage,blockandword. In invoice or contract processing, uncertain blocks can be routed to manual review. - Structured annotations and bbox extraction via
/v1/ocr, bulk processing via/v1/batch.
Pricing
Per the Mistral price list, OCR costs $4 per 1,000 pages, and $5 per 1,000 pages with annotations (Document AI). Compared with OCR 3 (around $2 per 1,000 pages) the per-page price has doubled – for high volumes it is worth calculating whether the new structural data justifies the premium. Prices per provider, as of October 2026.
EU availability and GDPR
- EU headquarters: French company subject to European data protection law.
- Mistral Cloud (Paris): native EU hosting option.
- Microsoft Foundry: Data Zone Standard EU, including the Frankfurt region (germanywestcentral), France and West Europe.
- AWS Bedrock and Google Vertex AI Model Garden: EU availability depends on model and region; Medium 3.5 and Small 4 are currently not listed there (see the compliance table above).
- Self-hosting: Small 4 (Apache 2.0) plus Medium 3.5 (modified MIT licence) and Large 3 (Mistral Research License) as open weights; no data transfer to third countries required.
Model lifecycle
- Leanstral 1.5 (2 July 2026): open-weights model (119B MoE, 6B active, Apache 2.0,
mistralai/Leanstral-1.5-119B-A6B) for proof engineering in Lean 4. According to Mistral, 100% on miniF2F and 587 of 672 PutnamBench problems. The free Labs API endpoint (labs-leanstral-1-5) was limited until 30 September 2026 per the changelog and is deprecated; self-hosting the weights remains possible. - OCR 4 (
mistral-ocr-4-0, released 23 June 2026): retired as of 30 September 2026 per the changelog; the successor is OCR 4.1 at the same price. - Magistral and Devstral 2: merged into Medium 3.5.
Other developments at Mistral
- Robostral Navigate (8 July 2026): first robotics model, 8B parameters for embodied navigation from natural-language instructions and a single RGB camera. According to Mistral 79.4% (val-seen) and 76.6% (val-unseen) on R2R-CE. Mistral has not announced licence or availability.
- Vibe: since late May 2026 Le Chat is called Vibe – an agent for work and code with Work Mode (connectors to Google Workspace, Outlook, SharePoint, Slack, GitHub) and Code Mode (remote coding, pull requests), powered by Medium 3.5.
- Funding: on 8 September 2026 Mistral announced a EUR 3 billion Series D led by Samsung Electronics at a post-money valuation of over EUR 21 billion (per Mistral). No new EU data centre locations were named.
Integration with CompanyGPT
All Mistral models are integrated into CompanyGPT and can be operated fully GDPR-compliantly in the EU – via Mistral Cloud (Paris), Microsoft Foundry (Data Zone Standard EU) or self-hosted; on AWS Bedrock and Google Vertex AI, EU region availability depends on the model version (see the compliance table above).
Our recommendation
We recommend a two-track strategy: Mistral Medium 3.5 for enterprise agents, coding and reasoning – thanks to toggle reasoning, 77.6% SWE-Bench Verified (per Mistral) and EU data residency including Copilot Studio integration. Mistral Small 4 remains the cost-efficient choice for standard workloads thanks to its Apache 2.0 licence and MoE efficiency. For open-weight flagship models there is Mistral Large 3; for specialised workloads Voxtral (audio), Mistral OCR 4.1 (documents), the Ministral 3 family (edge) and Leanstral 1.5 (formal proofs, self-hosting).