Claude 3.5 Sonnet (20240620): benchmark results
June 2024 Claude 3.5 Sonnet snapshot, kept separate from later Claude 3.5 Sonnet rows. Provider: Anthropic. Released 2024-06-20. Access: API.
Aggregate Rating 1656 ± 6, rank #299 of 1656 rated models, from 1561 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| BankMathBench (Korean) - Advanced | 15.8 | Exact-match accuracy (%) on the 399 Korean advanced (one pro | 100 |
| BankMathBench (Korean) - Basic | 69.2 | Exact-match accuracy (%) on the 559 Korean basic (single dep | 100 |
| CRMArena - BRI (Act) | 26.9 | Exact match (%) | 100 |
| CRMArena - KQA (Function Calling Native) | 40.5 | F1 (%) | 100 |
| CRMArena - NCR (Act) | 78.5 | Exact match (%) | 100 |
| CRMArena - NED (Act) | 51.5 | Exact match (%) | 100 |
| CRMArena - NED (Function Calling Native) | 52.3 | Exact match (%) | 100 |
| CRMArena - NED (ReAct) | 52.3 | Exact match (%) | 100 |
| CRMArena - Overall (Act) | 37.4 | Mean task score (%) | 100 |
| CRMArena - TCU (Function Calling Native) | 82.3 | Exact match (%) | 100 |
| Clembench Multimodal v1.6.5 | 80.77 | clemscore (imported) | 100 |
| HELM EWoK - Material Dynamics | 90.9 | EM | 100 |
Interactive version: theaggregate.ai/model?slug=claude-3-5-sonnet-20240620 · How It Works · Data refreshed daily, snapshot 2026-10-20.