AI Model Profile · Mistral AI
Mistral Large 4
Mistral Large 4 is an AI model developed by Mistral AI (France), released October 6, 2026. It has approximately 1.1T parameters. It was trained on about 1.2×1025 FLOP of compute. Its weights are openly released.
Open weights · Epoch confidence: Confident · Source: Epoch AI · Updated 2026-11-16
Specifications
| Developer | Mistral AI |
| Country | France |
| Released | October 6, 2026 |
| Domain | Language,Multimodal,Vision |
| Parameters | 1.1T |
| Training compute | 1.2×1025 FLOP |
| Training hardware | 3,800× NVIDIA GB200 |
| Training power draw | 8.8 MW |
| Weights | Open |
| Access | Open weights (restricted use) |
All figures from Epoch AI's Notable AI Models database (CC BY 4.0). Blank fields are undisclosed by the developer — never estimated by Affärslivet.
Running an open-weight model like Mistral Large 4 on your own infrastructure is an integration project as much as a model choice — our ranking of AI implementation partners compares who delivers that in Europe.
Benchmarks
Performance on Epoch AI's reproducible evaluations (higher is better; scores are the best across scorers). These are independent runs, not developer-reported numbers.
Source: Epoch AI Benchmarking Hub. GPQA Diamond = graduate-level science Q&A; SWE-Bench Verified = real software-engineering tasks; FrontierMath = research-level maths; MATH Level 5 = hardest competition maths; AIME = olympiad maths; SimpleQA = factual accuracy.
Frequently asked
- Who developed Mistral Large 4?
- Mistral Large 4 was developed by Mistral AI (France) and released on October 6, 2026.
- How big is Mistral Large 4?
- Mistral Large 4 has approximately 1.1T parameters and was trained on about 1.2×10^25 FLOP of compute. Source: Epoch AI.
- Is Mistral Large 4 open source?
- Mistral Large 4 is released with open weights (Yes), meaning its trained parameters are publicly available.
- How does Mistral Large 4 perform on benchmarks?
- On Epoch AI's evaluations, Mistral Large 4 scores 90.2% on GPQA Diamond, 22.4% on SimpleQA, 99.4% on AIME (OTIS Mock). Source: Epoch AI Benchmarking Hub.
Sources
Cite this model profile
Free to cite, quote and reference under CC BY 4.0 — with attribution to Affärslivet. Writing an article or answer? Reference this model profile as:
APA
Affärslivet Research. (2026). Mistral Large 4 — specifications & benchmarks. Affärslivet. https://xn--affrslivet-s5a.com/en/ai/models/mistral-large-4
MLA
"Mistral Large 4 — specifications & benchmarks." Affärslivet, 2026-11-16, xn--affrslivet-s5a.com/en/ai/models/mistral-large-4.
Source: Affärslivet — xn--affrslivet-s5a.com/en/ai/models/mistral-large-4. Attribute to Affärslivet when citing or linking.