GeoBrowse - Level 1 Country (Direct): leaderboard
GeoBrowse (2026) Level 1: models compose weak visual cues in a street scene into a location. This board scores the country-level targets with the model answering directly without tools; accuracy in percent from an LLM judge, mean of three runs, higher is better.
Metric: Accuracy (%) on the country-level targets of GeoBrowse Level 1 (199 street-scene instances whose location must be composed from weak, fragmented visual cues), judged correct by an LLM judge with the Humanity's Last Exam judging prompt, mean of 3 runs, under direct inference: the model answers in a single pass without tool calls; higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 12 models tracked. Updated 7 Oct 2026.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3 Pro | 53.3 | #145 |
| 2 | Claude Opus 4.5 | 43.4 | #160 |
| 3 | Gemini 2.5 Pro | 41.3 | #276 |
| 4 | Gemini 2.5 Flash | 40.2 | #463 |
| 5 | GPT-5 | 38 | #178 |
| 6 | GPT-4o | 35.7 | #751 |
| 7 | GPT-5.2 | 30.4 | #185 |
| 8 | GPT-4.1 | 30.4 | #505 |
| 9 | Llama 3.2 90B | 19.6 | #1755 |
| 10 | Qwen 2.5 VL 32B | 17.3 | #941 |
| 11 | Qwen 3 VL 32B | 16.3 | #622 |
| 12 | Qwen 2.5 VL 7B | 11.1 | #1862 |
No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.
Interactive version: theaggregate.ai/benchmark?slug=geobrowse-level-1-country-direct · How It Works · Data refreshed daily, snapshot 2026-11-21.