Company
Mistral AI
Every release, deal and policy move from Mistral AI — updated daily.
TopicsAll topics
-
Mistral Large 4: Europe’s Trillion-Parameter MoE Model
Mistral Large 4 launches as a 1.05T-parameter multimodal MoE model with 1M context. Discover its impact on enterprise AI, open-weight alternatives, and infrastructure challenges. Explore the guide.
By
Timeline
Release
Mistral Large 4
Open-weight MoE multimodal model, public preview (weights coming soon).
docs.mistral.ai (source, opens in a new tab)Funding
Mistral raises €3B Series D at ~€21B valuation
reuters.com (source, opens in a new tab)Partnership
Microsoft commits multiple billions to Mistral’s European GPU capacity
reuters.com (source, opens in a new tab)Release
Mistral Medium 3.5
docs.mistral.ai (source, opens in a new tab)Release
Mistral Small 4
docs.mistral.ai (source, opens in a new tab)Acquisition
Mistral AI acquires Koyeb, its first acquisition, for Mistral Compute
koyeb.com (source, opens in a new tab)Release
Devstral 2
docs.mistral.ai (source, opens in a new tab)Release
Mistral Large 3 & Ministral 3
docs.mistral.ai (source, opens in a new tab)
Sources: company announcements & release notes, press reports.
Fact box
- HQ
- Paris, France
- Founded
- 2023
- CEO
- Arthur Mensch
- Valuation
- ~€21B / ~$24.4B (Series D, Sept 2026)
- Total funding
- ~€6B (New York Times, 8 Sep 2026, incl. Series D)
- Core products
- Mistral Large · Le Chat · Codestral · Mistral Compute
- Latest round
- €3B Series D · Sep 2026
- Stories
- 44
Sources: mistral.ai, reuters.com, nytimes.com
Latest on X
Mistral AI · @MistralAI
Today we’re announcing a new collaboration with @nvidia to bring AI to industrial engineering: – faster simulation – open benchmarks and supporting datasets for physics AI – evaluations for agentic AI in CAE and EDA More here:
Introducing Voxtral 3: state-of-the-art Transcribe and Speech models powering Mistral’s new voice agents API. 🧵
Turns out le Chonk is into rocket science. Mistral Large 4 scoring well in engineering and astronomy on Ulam’s private benchmark, STEMBench:
What can large-scale AI models bring to the growing complexity of particle physics? @CERN and Mistral join forces to explore this question. The ambition: a single Large Physics Model, trained on billions of simulated collisions.
Welcome to Mistral, New Material. Let’s bring design to mission-critical AI.
Editor's picks
Deals & funding
Mistral AI · rounds, valuations, partners, acquisitions · $6.75B raised in tracked primary rounds · latest 8 of 9
Funding
Mistral raises €3B Series D at ~€21B valuation
reuters.com (source, opens in a new tab)Partnership
Microsoft commits multiple billions to Mistral’s European GPU capacity
reuters.com (source, opens in a new tab)Acquisition
Mistral AI acquires Koyeb, its first acquisition, for Mistral Compute
koyeb.com (source, opens in a new tab)Funding
Mistral AI raises €1.7B Series C led by ASML
Post-money valuation €11.7B (≈$13.7B); ASML invests €1.3B. USD amounts converted at announcement.
mistral.ai (source, opens in a new tab)Partnership
Mistral Compute: European AI cloud with NVIDIA (18,000 Grace Blackwell chips)
Announced at VivaTech with NVIDIA; sovereign GPU cloud in Essonne, France, initially 18,000 NVIDIA Grace Blackwell superchips.
mistral.ai (source, opens in a new tab)Funding
Mistral AI raises €600M Series B at ~$6B valuation
techcrunch.com (source, opens in a new tab)Partnership
Microsoft partnership: Mistral Large on Azure, €15M investment
azure.microsoft.com (source, opens in a new tab)Funding
Mistral AI raises €385M Series A at ~$2B valuation
techcrunch.com (source, opens in a new tab)
All 9 rounds & deals as a table · 5 funding, 3 partnerships, 1 acquisitions
| Date | Event | Type | Counterparty / round | Amount | Valuation | Source |
|---|---|---|---|---|---|---|
| Sep 8, 2026 | Mistral raises €3B Series D at ~€21B valuation Valuation ~€21B (≈$24.4B). Confirmed by Reuters and TechCrunch. | Funding | Series D | €3B | $24.4B | reuters.com |
| Jul 21, 2026 | Microsoft commits multiple billions to Mistral’s European GPU capacity Microsoft will use Mistral Compute capacity in Europe and add Mistral models to Azure, Foundry and Copilot Studio; not a new equity investment (Reuters). | Partnership · compute | with Microsoft | — | — | reuters.com |
| Feb 17, 2026 | Mistral AI acquires Koyeb, its first acquisition, for Mistral Compute Paris-based serverless platform; strengthens inference, sandboxes and on-prem deployment for Mistral Compute (TechCrunch). | Acquisition | acquires Koyeb | — | — | koyeb.com |
| Sep 9, 2025 | Mistral AI raises €1.7B Series C led by ASML Post-money valuation €11.7B (≈$13.7B); ASML invests €1.3B. USD amounts converted at announcement. | Funding | Series C | €1.7B | $13.7B | mistral.ai |
| Jun 11, 2025 | Mistral Compute: European AI cloud with NVIDIA (18,000 Grace Blackwell chips) Announced at VivaTech with NVIDIA; sovereign GPU cloud in Essonne, France, initially 18,000 NVIDIA Grace Blackwell superchips. | Partnership · compute | with NVIDIA | — | — | mistral.ai |
| Jun 11, 2024 | Mistral AI raises €600M Series B at ~$6B valuation €600M round led by General Catalyst: €468M equity plus €132M debt; valuation about $6B. | Funding | Series B | €600M | $6B | techcrunch.com |
| Feb 26, 2024 | Microsoft partnership: Mistral Large on Azure, €15M investment Multi-year deal: Mistral models on Azure AI, supercomputing for training. Microsoft's €15M (~$16M) minority investment converts into equity in a later round (reported by Reuters/FT). | Partnership · investment | with Microsoft | $16M | — | azure.microsoft.com |
| Dec 11, 2023 | Mistral AI raises €385M Series A at ~$2B valuation €385M (≈$415M) Series A led by Andreessen Horowitz; reported valuation about $2B. | Funding | Series A | €385M | $2B | techcrunch.com |
| Jun 13, 2023 | Mistral AI raises €105M seed round €105M seed (≈$113M) led by Lightspeed, weeks after founding by ex-DeepMind and Meta researchers; reported valuation about $260M. | Funding | Seed | €105M | $260M | techcrunch.com |
Sources: company announcements & release notes, press reports.
-
Mistral Large 4 Le Chonk: 1.05T Multimodal Model Public Preview
Mistral Large 4 (Le Chonk) launches as a 1.05T MoE model with 1M context and 256K output. Access…
-
Mistral Large 4: Europe’s Top Open-Weight Model
Mistral Large 4 launches as a 1T-parameter MoE model optimized for enterprise coding and cybersecurity. See how it…
-
Mistral AI Raises $830M for NVIDIA Blackwell Data Center
Mistral AI secured $830 million in debt to build a 44MW data center near Paris with 18,000 NVIDIA…
More stories
-
Safety Probes Follow Scaling Laws: Monitor Quality Can Be Predicted in Advance
Across 20 open base models and six safety tasks, probe error follows two simple laws. Fitted on small models, they predict larger ones and pick the…
-
AI Agents That Learn the Facts Still Deliver Only 60–70% of the Business Value
FDE-Gym’s 95 real delivery tasks: Claude Opus 5.5 leads at 0.56, but even models that gather 72–85% of key facts deliver just 0.60–0.70 of the reference…
-
Private LLM Servers: Solving Enterprise Data Privacy Issues
Explore how private LLM servers help organizations run models like Llama and Mistral securely without data leaks. Learn about infrastructure challenges and managed solutions. Discover the…
-
Mistral Medium 3.5 in Copilot Studio: Multi-Model Enterprise AI
Microsoft adds Mistral Medium 3.5 to Copilot Studio, letting enterprises build agents with European models inside Microsoft 365. Move beyond OpenAI lock-in for RAG and agent…
-
Mistral AI: Enterprise Data Sovereignty with On-Prem LLMs
Mistral AI offers open-weight models like Mixtral that run inside enterprise data centers, cutting cloud costs and meeting strict data privacy rules. Learn how to deploy…
-
Open-Weight Models Fracture Proprietary LLM Monopoly
Open-weight models from Meta, Google, and Mistral are commoditizing base intelligence and reshaping AI economics. Learn how enterprises gain deployment flexibility and lower TCO. Explore the…
-
Agentic Search: Moving Beyond Basic RAG for Better AI Accuracy
Explore how Agentic Search from Mistral, Cohere, and LlamaIndex turns LLMs into autonomous researchers. Understand accuracy gains, token costs, and production tradeoffs. Discover the guide.
-
Mistral AI: Sovereign Alternative for Regulated Enterprises
Mistral AI pairs open-weight models with GDPR-compliant hosting to give enterprises a data-sovereign hedge against US vendors. See how Mixture-of-Experts designs cut TCO for finance, healthcare,…
-
Open-Weights Models Shift AI Focus to Infrastructure & Licenses
Open-weights models like Llama 3 and Mixtral now rival closed systems. Enterprises gain lower TCO, data control, and hybrid routing options while navigating complex licensing. Explore…
-
Scaleup Europe Fund: $5B Boost for EU AI Sovereignty
The EU’s $5 billion Scaleup Europe Fund closes late-stage gaps for AI firms like Mistral AI. Explore how it supports sovereign compute and prevents talent drain…
-
Mistral OCR 4: Self-Hosted OCR Engine for Private Enterprise AI
Mistral OCR 4 delivers a self-hosted Document AI engine supporting 170 languages. Run accurate extraction on-premises to cut TCO, meet GDPR and HIPAA requirements, and power…
-
Mistral AI Enterprise Registry: Agents, Judges & Governance
Mistral AI is building enterprise-grade orchestration with its AI Studio registry for agents, tools, workflows, and judges. Discover how it delivers compliance and auditability for regulated…
Models & specs
Mistral AI · context window, max output, API price, modalities, knowledge cutoff, license
| Model | Released | Context | Max output | Price in / out (per 1M) | Modalities | Knowledge cutoff | License / access | Spec source |
|---|---|---|---|---|---|---|---|---|
| Mistral Large 4 | Oct 6, 2026 | 1M | — | $1.36 / $4.18 cached $0.14 | Text, image in → text out | — | Open weights (public preview) 1.05T total / 52B active (MoE) + 1.6B vision encoder | docs.mistral.ai |
| Mistral Medium 3.5 | Apr 28, 2026 | 256K | — | $1.5 / $7.5 | Text, image in → text out | — | Open weights, modified MIT | docs.mistral.ai |
| Mistral Small 4 | Mar 16, 2026 | 256K | — | $0.15 / $0.6 | Text, image in → text out | — | Open weights, Apache 2.0 119B total / 6.5B active (MoE) | docs.mistral.ai |
| Devstral 2 | Dec 9, 2025 | 256K | — | open weights | Text in → text out | — | Open weights, modified MIT | docs.mistral.ai |
| Mistral Large 3 & Ministral 3 Mistral Large 3; Ministral 3 3B, 8B, 14B | Dec 2, 2025 | 256K | — | $0.5 / $1.5 Ministral 3 14B: $0.20/$0.20. | Text, image in → text out | — | Open weights, Apache 2.0 675B total / 41B active (MoE) | docs.mistral.ai |
| Magistral Magistral Medium, Magistral Small | Jun 10, 2025 | 40K | — | open weights | Text in → text out | — | Medium: proprietary · Small: open weights, Apache 2.0 | docs.mistral.ai |
| Mistral Medium 3 | May 7, 2025 | 128K | — | — | Text, image in → text out | — | Proprietary (API) | docs.mistral.ai |
| Mistral Small 3 | Jan 30, 2025 | 32K | — | open weights | Text in → text out | — | Open weights, Apache 2.0 24B dense | docs.mistral.ai |
| Mistral Large 2 | Jul 24, 2024 | 128K | — | open weights | Text in → text out | — | Open weights, Mistral Research License (MRL) 123B dense | docs.mistral.ai |
| Codestral | May 29, 2024 | 32K | — | open weights | Text in → text out | — | Open weights, Mistral Non-Production License (MNPL) 22B dense | docs.mistral.ai |
Sources: company announcements & release notes. Prices = standard API list price in USD per 1M tokens at time of release or current docs; batch is typically 50% off.