SpatialBench V1 Verified
Science
SpatialBench V1 verified subset of 115 spatial-transcriptomics tasks from the original 159-task release. Published per-model and per-harness pass rate.
Top models (higher is better)
| Model | Score |
|---|---|
| Grok 4.6 | 77.4 |
| DeepSeek V4.1 Flash | 75.4 |
| GPT-5.6 Sol | 74.5 |
| Claude Opus 5 | 71.0 |
| Grok 4.5 | 71.0 |
| GPT-5.6 Terra | 70.4 |
| GPT-6 Astra | 70.1 |
| Sonnet 5 | 69.3 |
| Opus 4.8 | 69.0 |
| Gemini 3.5 Flash | 67.0 |
| GPT-5.5 | 65.2 |
| Opus 4.7 | 64.1 |
| Kimi K3 | 61.5 |
| GPT-5.6 Luna | 60.6 |