Pixtral Large: benchmark results
Mistral's 124B frontier multimodal model (November 2024) built on Mistral Large 2, with weights released under the research-only MRL license. Provider: Mistral. Released 2024-11-18. Access: Open.
Aggregate Rating 1554 ± 15, rank #1654 of 10916 rated models, from 68 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Visual Budgeting Transfer - Vis. Haystacks - Full resolution | 78.9 | Accuracy (%) | 100 |
| Visual Budgeting Transfer - Vis. Haystacks - mmVB | 77.4 | Accuracy (%) | 100 |
| LLM Stats (ChartQA) | 88.1 | Score (%) | 73.1 |
| HAERAE-Vision - Entertainment/Arts | 11.54 | Checklist Score (%; original under-specified queries, GPT-5- | 63.6 |
| LLM Stats (MathVista) | 69.4 | Score (%) | 60.5 |
| HAERAE-Vision - Daily Life | 24.16 | Checklist Score (%; original under-specified queries, GPT-5- | 59.1 |
| HAERAE-Vision - Gaming | 11.33 | Checklist Score (%; original under-specified queries, GPT-5- | 59.1 |
| HAERAE-Vision - Transportation | 21.93 | Checklist Score (%; original under-specified queries, GPT-5- | 59.1 |
| LLM Stats (DocVQA) | 93.3 | Score (%) | 58.9 |
| HAERAE-Vision - Coding/Development | 22.64 | Checklist Score (%; original under-specified queries, GPT-5- | 56.8 |
| HAERAE-Vision | 20.1 | Checklist Score (%; original under-specified queries, GPT-5- | 54.5 |
| HAERAE-Vision - IT/Computer | 19.09 | Checklist Score (%; original under-specified queries, GPT-5- | 54.5 |
Interactive version: theaggregate.ai/model?slug=pixtral-large · How It Works · Data refreshed daily, snapshot 2026-11-08.