Claude 4: benchmark results
Provider: Anthropic. Access: API.
Aggregate Rating 1670 ± 48, rank #505 of 10916 rated models, from 6 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| WMT25 English-Maasai chrF++ | 26.1 | chrF++ | 96.2 |
| WMT25 English-Maasai ESA | 7.7 | ESA Human Score | 94.1 |
| WMT25 English-Bhojpuri chrF++ | 36.7 | chrF++ | 85.3 |
| WMT25 English-Bhojpuri ESA | 83.2 | ESA Human Score | 82.4 |
| GosuEvals | 67.2 | Score (%) | 67.9 |
| DiscoX | 54.04 | Overall Score | 62.5 |
Interactive version: theaggregate.ai/model?slug=claude-4 · How It Works · Data refreshed daily, snapshot 2026-11-08.