HardGeoBench <=10mi
Multimodal · 2025-05-08
HardGeoBench percentage of private-photo geolocation predictions within 10 miles of the true location. Scores are pass@1 percentages from API default settings with a simple prompt, no sampling, and no web search or tools.
Top models (higher is better)
| Model | Score |
|---|---|
| Gemini 3.5 Flash | 45.0 |
| Claude Fable 5 | 41.3 |
| Gemini 3 Flash Preview | 38.8 |
| Gemini 3.1 Pro Preview | 38.8 |
| Gemini 2.5 Pro Preview 06-05 | 37.5 |
| Gemini 2.5 Pro Experimental 03-25 | 35.0 |
| o3-pro | 35.0 |
| Gemini 3 Pro Preview | 35.0 |
| o3 | 33.8 |
| GPT-5.1 | 33.8 |
| Opus 4.6 | 32.5 |
| ChatGPT-4o Latest (source-unspecified snapshot) | 31.3 |
| GPT-5 | 31.3 |
| GPT-5 (2025-08-07) | 31.3 |
| GPT-4.1 | 30.0 |
| Opus 4.5 | 30.0 |
| Kimi K2.5 | 30.0 |
| GPT-5.6 Sol | 30.0 |
| GPT-5.5 | 28.8 |
| Gemini 2.5 Flash Preview 05-20 | 27.5 |
| Gemini 2.5 Flash Preview (09-2025) | 27.5 |
| Sonnet 4.6 | 27.5 |
| Gemini 3.1 Flash-Lite Preview | 27.5 |
| GPT-5.4 | 27.5 |
| Opus 4.7 | 27.5 |
| Opus 4.8 | 27.5 |
| Gemini 2.0 Flash | 26.3 |
| GPT-5.6 Terra | 25.0 |
| GPT-5.2 (2025-12-11) | 23.8 |
| Gemini 2.0 Flash-Lite | 22.5 |
| O4 Mini | 21.3 |
| GPT-5 Mini | 21.3 |
| Sonnet 4.5 | 21.3 |
| Haiku 3.5 | 20.0 |
| Sonnet 5 | 20.0 |
| Grok 4 | 18.8 |
| GPT-5.6 Luna | 18.8 |
| Grok 4.5 | 17.5 |
| Gemini 1.5 Flash | 16.3 |
| Gemma 4 31B IT | 15.0 |