Anthropic · LIVEBENCH 2026-06-25

Claude Sonnet 5 xHigh Effort

Published benchmark results for this specific model configuration.

claude-sonnet-5-xhigh-effort

Overall score76.0 / 100

Cost / successful task$0.505USD · benchmark workload

Weight accessNot reported

Where this configuration performs

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

Reasoning88.7

Coding80.7

Agentic Coding59.4

Mathematics92.9

Data Analysis71.7

Language75.0

Instruction Following63.9

Task-level results

theory of mind80.8zebra puzzle88.0spatial100.0logic with navigation86.0code generation83.1code completion78.3javascript68.2typescript50.0python60.0AMPS Hard98.0integrals with game88.0math comp95.1olympiad90.7consecutive events74.7tablejoin42.4tablereformat98.0connections95.5plot unscrambling53.4typos76.0paraphrase61.3simplify60.6story generation62.3summarize71.2

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