Claude Sonnet 5.5 (xHigh): benchmark results

Provider: Anthropic. Released 2026-09-28. Access: API.

Unified ELO 1929 ± 25, rank #30 of 2059 rated models, from 30 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
LiveBench Code Generation92.96Score98.5
LiveBench Integrals With Game100Score96.9
Epoch AI - Critpt31.14Score96.5
LiveBench Consecutive Events90.94Score95.4
CursorBench 4.053.1Score (%)95
LiveBench Simplify72.37Score93.8
Bug Hunt Bench - LMS20Planted Bugs Fixed (out of 60)91.1
LLM2014 Logic 2026-0966.5Median Score90.9
LiveBench Code Completion84.78Score90.8
Bug Hunt Bench36Planted Bugs Fixed (out of 105)90.1
LiveBench Zebra Puzzle100Score86.2
Bug Hunt Bench - VS Code Extension16Planted Bugs Fixed (out of 45)84.9

Interactive version: theaggregate.ai/model?slug=claude-sonnet-5-5-xhigh · How It Works · Data refreshed daily, snapshot 2026-09-30.