Claude 3.7: benchmark results
Provider: Anthropic. Released 2025-02-24. Access: API.
Unified ELO 1678 ± 31, rank #377 of 2656 rated models, from 18 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| EduGuardBench | 77 | RFS (self-reported) | 100 |
| EComStage - Query Rewrite | 81.99 | Task Score (%) | 96.9 |
| EComStage - RAG-QA | 69.38 | Task Score (%) | 78.1 |
| GosuEvals | 70.6 | Score (%) | 77.8 |
| EComStage - Intent Recognition | 88 | Task Score (%) | 75 |
| EComStage - Query Match | 98.39 | Task Score (%) | 70.3 |
| KORGym - Puzzle | 0.55 | Score | 61.1 |
| EComStage - Scenario Route | 82.32 | Task Score (%) | 59.4 |
| EComStage | 81.89 | Task Score (%) | 53.1 |
| KORGym | 0.35 | Score | 44.4 |
| KORGym - Mathematical and Logical | 0.26 | Score | 44.4 |
| KORGym - Control and Interaction | 0.25 | Score | 41.7 |
Interactive version: theaggregate.ai/model?slug=claude-3-7 · How It Works · Data refreshed daily, snapshot 2026-09-19.