Mistral Medium 3.5: benchmark results
Mistral's open 128B multimodal Medium 3.5 model unifying chat, reasoning, and coding. Provider: Mistral. Released 2026-04-28. Access: Open.
Aggregate Rating 1631 ± 7, rank #353 of 1656 rated models, from 590 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| SEA-HELM (Filipino) - SEA-Safeguard | 79.65 | Normalized Score | 100 |
| Nejumi 4 - jaster (2-shot) - JSICK | 87 | Exact match (%) | 98.6 |
| SEA-HELM (Vietnamese) - Question Answering | 82.49 | Normalized Score | 96.5 |
| SEA-HELM (Vietnamese) - Summarization | 18.14 | Normalized Score | 96.5 |
| Nejumi 4 - BFCL - Live AST | 75 | Accuracy (%) | 95.1 |
| SEA-HELM (Vietnamese) - NLG | 55.73 | Normalized Score | 94.7 |
| SEA-HELM (Vietnamese) - NLU | 71.23 | Normalized Score | 94.7 |
| AA TAU-2 Bench | 94.15 | Accuracy (%) | 92 |
| Tau2-Bench Telecom | 94.15 | Success Rate (%) | 92 |
| Horangi 4 - Korean Hate Speech | 69 | Accuracy (%) | 91.8 |
| SEA-HELM (Filipino) - NLG | 56.98 | Normalized Score | 91.2 |
| SEA-HELM (Indonesian) - NLG | 55.78 | Normalized Score | 91.2 |
Interactive version: theaggregate.ai/model?slug=mistral-medium-3-5 · How It Works · Data refreshed daily, snapshot 2026-10-20.