Anthropic · LIVEBENCH 2026-06-25

Claude Sonnet 5.5 Max Effort

Published benchmark results for this specific model configuration.

claude-sonnet-5-5-max-effort

Overall score75.7 / 100

Cost / successful task$0.868USD · benchmark workload

Weight accessNot reported

Where this configuration performs

Category averages from the same release, on a 0–100 scale.

Reasoning91.6

Coding91.4

Agentic Coding56.3

Mathematics96.1

Data Analysis59.5

Language78.0

Instruction Following56.8

Task-level results

theory of mind86.5zebra puzzle100.0spatial98.0logic with navigation82.0code generation95.8code completion87.0javascript77.3typescript56.7python35.0AMPS Hard98.0integrals with game97.0math comp97.1olympiad92.5consecutive events40.6tablejoin39.8tablereformat98.0connections99.3plot unscrambling48.6typos86.0paraphrase48.2simplify50.9story generation69.9summarize58.2

Source files: Scores ↗ Categories ↗ Costs ↗ . Missing cost is unknown, never free. Compare configurations within the same release.