Skip to content

AnthropicFrontier

Claude Sonnet 5.5

Faster, cheaper companion to Opus 5.5.

Announced
Access
Public

Benchmark scores

TestScoreBest on the trackerHow it was measured
Epoch Capabilities Indexmany tests combined into one score166 pointsGPT-6.2 Astra 174 pointsIndependentlisted as "Claude Sonnet 5.5", CI 163.65 to 169.35Run 9 Nov 2026
AA Intelligence Index v5about ten tests in one score54.5 pointsClaude Fable 5.2 64.6 pointsIndependentv5.0.1; listed as Claude Sonnet 5.5 (Max, Default Fallback)Run 10 Nov 2026
FrontierSWE V2very long engineering jobs62%Claude Fable 5.2 70%Independentproximus harness, max effort, mean@5 over 34 tasks, 20-hour budget (best@5 71.7)
Terminal-Bench 4.1long command-line tasks64%Claude Opus 5.5 66%IndependentClaude Code harness, max effort, 330 trials, ±3 (95% CI)
APEX-Agentsprofessional office tasks76%Gemini 4 Carbon 84%Independentpass@1, max effort; listed as "Sonnet 5.5 [claude-sonnet-5-5]"
Humanity's Last Examexpert questions, no tools46%GPT-6 Astra 55%Independentleaderboard entry "Sonnet 5.5 (xhigh)" (claude-sonnet-5-5_xhigh)
ARC-AGI-3puzzle games, standard harnessTypical first-time human player: 100%24%Claude Fable 5.2 81%IndependentARC Prize config "Claude Sonnet 5.5 (XHigh)" (anthropic-claude-sonnet-5-5-xhigh), standard harness
FrontierMath T1–3hard maths problems89%Claude Fable 5.2 95%Independentmax reasoning effort (claude-sonnet-5-5_max)Run 29 Sept 2026
GPQA DiamondPhD-level science questionsPhD experts: 70%Saturated96%Gemini 4 Argon 96%Independentmax reasoning effort (claude-sonnet-5-5_max)Run 29 Sept 2026
FrontierMath T4research-level maths problemsSaturated80%GPT-6.1 Sol 100%Independentmax reasoning effort (claude-sonnet-5-5_max)Run 29 Sept 2026
ARC-AGI-2visual puzzles, model aloneAverage person: 60%Saturated84%GPT-6 Astra 95%IndependentARC Prize config "Claude Sonnet 5.5 (XHigh)" (anthropic-claude-sonnet-5-5-xhigh)
OTIS Mock AIMEcompetition mathsSaturated100%Top scoreIndependentmax reasoning effort (claude-sonnet-5-5_max)Run 29 Sept 2026

Compare every model on the same tests →

Sources