Workman et al. (scBench team)

SpatialBench: Can Agents Analyze Real-World Spatial Biology Data?

Real-world spatial biology analysis tasks across five technologies and seven task categories, with a 115-evaluation expert-verified subset. The live verified leaderboard has GPT-5.6 Sol at 74.5% with Codex and Opus 5 at 71.0% with Claude Code.

spatial-biologyspatial-transcriptomicstabularimagecode
Status
cataloged
Saturation
unsaturated
Reproduction
not yet
Published results
9 models

Editor's note

Real-world spatial biology analysis tasks across five technologies and seven task categories, with a 115-evaluation expert-verified subset. The live verified leaderboard has GPT-5.6 Sol at 74.5% with Codex and Opus 5 at 71.0% with Claude Code.

Published results

ModelScoreVariantProvenance
GPT-5.6 Sol74.5% accuracyCodex harness · verified subsetexternal
Opus 571.0% accuracyClaude Code harness · verified subsetexternal
Claude Opus 4.538.4% accuracydefaultpaper
GPT-5.234.0% accuracydefaultpaper
Claude Sonnet 4.528.3% accuracydefaultpaper
GPT-5.127.4% accuracydefaultpaper
Grok-4.124.7% accuracydefaultpaper
Grok-422.8% accuracydefaultpaper
Gemini 2.5 Pro20.1% accuracydefaultpaper

Sources

  • external benchmarks.bio — SpatialBench verified leaderboard
  • paper SpatialBench (arXiv:2512.21907)