Benchmarks.bio - SpatialBench-Long — leaderboard
Long-form Benchmarks.bio spatial transcriptomics tasks that require multi-step biological data analysis, tool use, and synthesis over larger assay contexts.
Metric: Pass Rate (%). Source: benchmarks.bio. Status: saturation imminent. 18 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5.6 Sol | 38.89 |
| 2 | GPT-5.6 Terra | 22.22 |
| 3 | Grok 4.5 | 19.44 |
| 4 | GPT-5.6 Luna | 18.06 |
| 5 | Claude Sonnet 5 | 12.5 |
| 6 | GPT-5.5 | 11.11 |
| 7 | Claude Opus 4.8 | 11.11 |
| 8 | Gemini 3.5 Flash | 11.11 |
| 9 | Claude Opus 4.6 | 9.72 |
| 10 | Claude Opus 4.7 | 8.33 |
| 11 | Claude Opus 4.8 (Max) | 8.33 |
| 12 | Grok 4.20 Beta (0309) (Reasoning) | 6.94 |
| 13 | GPT-5.4 | 5.56 |
| 14 | Kimi K2.6 | 5.56 |
| 15 | Claude Sonnet 4.6 | 4.17 |
Interactive version: theaggregate.ai/benchmark?slug=benchmarks-bio-spatialbench-long · How the rankings work · Data refreshed daily, snapshot 2026-07-22.