HealthBench ProBrowse 296

HealthBench Pro

Can the model produce useful work in a professional knowledge-work setting?

Version not specifiedExact reported variant
Knowledge Work5 ranked models5 reported values1 reportsHigher is better

Top rankings

One row per model, using its best reported score across effort settings.

5
RankModelBest scoreBest reported setting
1Muse Spark 1.1Meta59.3%xhigh effortMeta report1 value · 1 reportJul 9, 2026 · source
2Claude Opus 4.8Anthropic55.8%max effortMeta report1 value · 1 reportJul 9, 2026 · source
3Muse Spark 1.0Meta54.1%Reported configurationMeta report1 value · 1 reportJul 9, 2026 · source
4GPT-5.5OpenAI51.8%xhigh effortMeta report1 value · 1 reportJul 9, 2026 · source
5Gemini 3.1 ProGoogle41.6%high effortMeta report1 value · 1 reportJul 9, 2026 · source

Effort curve

Every sourced cost-linked effort value for this exact version. Lines connect complete sweeps only.

0
No cost-linked effort sweep for this version.
Definition and comparison boundarypublic methodology

Can the model produce useful work in a professional knowledge-work setting?

Professional analysis, document, finance, office, or domain-specific tasks. Higher is better. The value is the percentage reported in this lab's table.

Harness, tools, prompts, attempt count, and benchmark version can materially change the result. Compare within one reporting context.

525 data points; GPT-5.4 low-reasoning grader; length-normalized rubric score.

Method / source