GeoRC: leaderboard
Geolocation reasoning benchmark for vision-language models, measuring location inference, F1, and country-level accuracy.
Metric: F1 (imported). Source: talrejamohit03.github.io. Source: benchmarklist.com. Saturation forecast: Around April 2027. 12 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Best Expert | 53.92 | |
| 2 | GPT-4.1 | 42.3 | #248 |
| 3 | Gemini 2.5 Pro | 41.51 | #152 |
| 4 | Gemini 2.5 Flash | 41.3 | #245 |
| 5 | Gemini 3 Pro | 40.98 | #80 |
| 6 | GPT-5 | 40.56 | #98 |
| 7 | Qwen 2.5 VL 7B Instruct | 31.63 | #654 |
| 8 | Gemma 3 12B (IT) | 31.21 | #664 |
| 9 | Llama 3.2 11B Instruct | 25.86 | #1126 |
| 10 | Qwen 3 VL 8B Instruct | 23.81 | #420 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=georc · How It Works · Data refreshed daily, snapshot 2026-10-20.