GeoScope Study - GeoScope - OSR: leaderboard
GeoScope v1,OSR,400testquestions,native accuracy percent,GeoScope. Seeded20%development/80%test,configuration frozen. Direct wholeimage4Mpx,firsttoken optionlogit; numeric-only response options form separate condition. GeoScope uses fixedQwen3-4Bsynthesizer,greedy,horizon5,depth5 withnamedperceptionmodel. Exact rerun protocol retained,not a framework-named checkpoint. No constant-answer test-tie upperbound or cited external scores.
Metric: Score (%). Source: arxiv.org. Saturation forecast: Rough model projection: around 2026. 4 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen3-VL-8B-Instruct (GeoScope Study Run: method=GeoScope) | 44 | |
| 2 | Qwen3-VL-4B-Instruct (GeoScope Study Run: method=GeoScope) | 40.75 | |
| 3 | Qwen2.5-VL-7B-Instruct (GeoScope Study Run: method=GeoScope) | 40.25 | |
| 4 | llava-onevision-qwen2-7b-ov-hf (GeoScope Study Run: method=GeoScope) | 31 |
Interactive version: theaggregate.ai/benchmark?slug=geoscope-study-geoscope-osr · How It Works · Data refreshed daily, snapshot 2026-11-12.