Claude Sonnet 5.5

Where people rank it

From blind votes on LMArena, where people pick the better of two anonymous answers. See the full tables on the AI leaderboard.

On paper

Fixed-test rankings from the benchmarks tab: Reasoning #4, Speed #5, Value #14.

Strengths and weaknesses

Best for

• Everyday coding in Claude Code
• Running AI agents that use tools
• A fast, capable daily assistant

Every score we have

Each bar compares this model with the best result any model on this page has on the same test.

• LMArena text arena (human votes)1,474.9Arena score 1474.9 (95% CI 1467.2 to 1482.7); 6,831 votes; entry "claude-sonnet-5.5-xhigh"; position 36 of 416 on the full board
• LMArena creative writing1,465.1Arena score 1465.1 (95% CI 1448.9 to 1481.3); 1,532 votes; entry "claude-sonnet-5.5-xhigh"; position 22 of 414 on the full board
• LMArena WebDev (coding)1,767.6Arena score 1767.6 (95% CI 1754.4 to 1780.8); 3,182 votes; entry "Claude Sonnet 5.5 (xHigh)"; position 3 of 144 on the full board
• LMArena vision1,266.2Arena score 1266.2 (95% CI 1252.5 to 1279.9); 2,068 votes; entry "claude-sonnet-5.5-xhigh"; position 39 of 165 on the full board
• LMArena Agent Arena0.130IPS score 0.1299 (95% CI 0.0986 to 0.1611); 7,560 sessions; entry "Claude Sonnet 5.5 (Max)"; position 6 of 52 on the full board
• Artificial Analysis Intelligence Index5656.0 on the Artificial Analysis Intelligence Index v4.3.2 (AA shows 56), setting "Claude Sonnet 5.5 (max with fallback)"
• Humanity's Last Exam (Artificial Analysis)55.0%55.0% on Humanity's Last Exam (AA, 2,158 text-only questions), setting "Claude Sonnet 5.5 (max with fallback)"
• Terminal-Bench 4.0 (Artificial Analysis run)63.6%63.6% on Terminal-Bench 4.0 as run by AA (mini-swe-agent harness), setting "Claude Sonnet 5.5 (max with fallback)"
• Long-context reasoning (AA-LCR)82.7%82.7% on AA-LCR (long-context reasoning), setting "Claude Sonnet 5.5 (max with fallback)"
• AA-Omniscience, accuracy minus hallucination32.332.3 on the AA-Omniscience Index (knowledge accuracy minus hallucination, range -100 to 100), setting "Claude Sonnet 5.5 (max with fallback)"
• Output speed, tokens per second141 t/s141 output tokens per second (AA median), row "Claude Sonnet 5.5 (max with fallback)"
• Cost per Artificial Analysis test task$5.46$5.46 to run one AA Intelligence Index task, row "Claude Sonnet 5.5 (max with fallback)"
• Artificial Analysis Coding Agent Index68.468.4 on the AA Coding Agent Index v1.5 (DeepSWE v1.1, Terminal-Bench 4.0, SWE-Atlas-QnA), best listed setup "Claude Code - Sonnet 5.5 (max)"
• Terminal-Bench 4.0 leaderboard61.8%61.8% resolved on Terminal-Bench 4.0 (plus or minus 2.9, 95% CI) with Claude Code at max effort
• Humanity's Last Exam (Scale AI)46.1%46.12% on Scale AI's Humanity's Last Exam leaderboard (plus or minus 1.95), entry "Sonnet 5.5 (xhigh)"