GeoReader Spatial Suite - SPBench: leaderboard
GeoReader Spatial Suite - SPBench. Source reported spatial benchmark scores;† reproduced base-model runs and *author baseline retained with exact family/size. Trained reader checkpoints are distinct models, not backbone aliases. Metric: Accuracy (%). Primary published table; source configurations remain distinct.
Metric: Accuracy (%). Source: iris0329.github.io. Saturation forecast: Around January 2027. 20 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 3.5 9B | 50.7 | #367 |
| 2 | Qwen 2.5 VL 7B | 40.8 | #632 |
| 3 | Qwen 3.5 4B | 35.5 | #463 |
No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #7 Gemini 4 Argon.
Interactive version: theaggregate.ai/benchmark?slug=georeader-spatial-suite-spbench · How It Works · Data refreshed daily, snapshot 2026-10-16.