AstaBenchBrowse 296

AstaBench

Can the model reason accurately about difficult scientific material?

Version not specifiedExact reported variant
Science7 ranked models7 reported values1 reportsHigher is better

Top rankings

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

7
RankModelBest scoreBest reported setting
1Claude Opus 4.7Anthropic58.0%max effort$3.54 / task1 value · 1 reportApr 30, 2026 · source
2Claude Opus 4.6Anthropic55.3%max effortAllen AI report1 value · 1 reportApr 30, 2026 · source
3Claude Sonnet 4.6Anthropic54.5%max effortAllen AI report1 value · 1 reportApr 30, 2026 · source
4Asta v0Allen AI53.0%Reported configurationAllen AI report1 value · 1 reportApr 30, 2026 · source
5GPT-5.5OpenAI52.9%xhigh effort$1.61 / task1 value · 1 reportApr 30, 2026 · source
6Gemini 3.1 Pro PreviewGoogle49.6%high effortAllen AI report1 value · 1 reportApr 30, 2026 · source
7GPT-5.4OpenAI46.5%xhigh effortAllen AI report1 value · 1 reportApr 30, 2026 · source

Effort curve

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

2
No cost-linked effort sweep for this version.
Definition and comparison boundarypublic methodology

Can the model reason accurately about difficult scientific material?

Scientific knowledge and reasoning under the benchmark's published question set. 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.

ReAct agent framework. The source directly reports observed per-problem cost only for Claude Opus 4.7 ($3.54) and GPT-5.5 ($1.61).

Method / source