MILU: leaderboard
Multilingual knowledge-and-reasoning evaluation reported in Anthropic's Claude Opus 4.8 system card.
Metric: Score (self-reported). Source: benchmarklist.com. Status: years away from saturation. 13 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3.1 Pro (Preview) | 93.6 |
| 2 | Claude Fable 5.1 (Max) | 93 |
| 3 | Claude Fable 5 (Max) | 92.9 |
| 4 | Claude Mythos 5 | 92.9 |
| 5 | Claude Mythos Preview | 92.7 |
| 6 | Claude Opus 5 | 92.1 |
| 7 | Claude Opus 5 (Max) | 92.1 |
| 8 | Claude Opus 4.8 | 91.1 |
| 9 | GPT-5.4 | 90.6 |
| 10 | Claude Opus 4.7 | 89.9 |
| 11 | Claude Sonnet 4.6 | 89.4 |
| 12 | Claude Sonnet 5 | 89.3 |
| 13 | Claude Sonnet 5 (Max) | 89.3 |
Interactive version: theaggregate.ai/benchmark?slug=milu · How It Works · Data refreshed daily, snapshot 2026-09-05.