Claude Sonnet 5.5 (Low): benchmark results
Provider: Anthropic. Released 2026-09-28. Access: API.
Aggregate Rating 1804 ± 23, rank #128 of 2145 rated models, from 16 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| CMDB-1500 (Image) | 92.67 | Accuracy (%), fixed denominator of 300 questions (invalid an | 100 |
| Conceptual Reasoning Index - Argument Evaluation (LMCA) | 64.54 | Chance-Corrected Score (0-100) | 94.6 |
| Conceptual Reasoning Index - Consistency (ACCoRD) | 82.92 | Chance-Corrected Score (0-100) | 94.3 |
| Conceptual Reasoning Index | 73.08 | Chance-Corrected Score (0-100) | 90.9 |
| CMDB-1500 (Text) | 78 | Accuracy (%), fixed denominator of 1,200 questions (invalid | 77.8 |
| Conceptual Reasoning Index - Decision Theory (DTBench) | 88.88 | Chance-Corrected Score (0-100) | 75.9 |
| Epoch AI - Critpt | 11.43 | Score | 56.1 |
| Bug Hunt Bench - LMS | 12.3 | Planted Bugs Fixed (out of 60) | 53.5 |
| SelfBench - supabase/supabase | 72.22 | Task success (%) on 36 tasks rebuilt from the repository's m | 45.5 |
| Bug Hunt Bench | 19 | Planted Bugs Fixed (out of 105) | 41.6 |
| CursorBench 4.0 | 35.8 | Score (%) | 35.4 |
| DuelLab Overall | 27.1 | GameBench2 normalized score (0–100) | 33.8 |
Interactive version: theaggregate.ai/model?slug=claude-sonnet-5-5-low · How It Works · Data refreshed daily, snapshot 2026-10-16.