Pixtral-12B-2409: benchmark results
Provider: Mistral. Released 2024-09-11. Access: Open.
Unified ELO 1407 ± 1, rank #1107 of 1392 rated models, from 23 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| KOFFVQA - Recognition | 65 | Score (%) | 45.1 |
| KOFFVQA - Graph and Chart Understanding | 56 | Score (%) | 44.4 |
| TextClass Benchmark | 1490.38 | Meta-Elo (self-reported) | 38.3 |
| Clembench Multimodal v1.6.5 | 28.64 | clemscore (self-reported) | 36.8 |
| KOFFVQA - Commonsense Reasoning | 65.11 | Score (%) | 30.9 |
| KOFFVQA - Korean Recognition | 10 | Score (%) | 30.9 |
| KOFFVQA - Object Attributes | 61.67 | Score (%) | 30.9 |
| KOFFVQA - Table Understanding | 23.67 | Score (%) | 27.2 |
| KOFFVQA | 44.62 | Overall Score (%) | 25.9 |
| KOFFVQA - Relationship | 45.67 | Score (%) | 25.3 |
| KOFFVQA - Korean OCR | 10 | Score (%) | 19.1 |
| U-MATH - Differential Calculus | 16.82 | Accuracy (%) | 18.2 |
Interactive version: theaggregate.ai/model?slug=pixtral-12b-2409 · How It Works · Data refreshed daily, snapshot 2026-09-05.