multimodal benchmark · included in ECI
GeoBench
GeoBench benchmark (score column: ACW Country %).
Results
5
Random baseline
0.0%
Score ceiling
100%
Released
1 Mar 2025
Score by model release date
Best score by organisation
Leaderboard
| # | Model | Score | Relative | Setting | Released | Run | Evidence |
|---|---|---|---|---|---|---|---|
| 1 | Qwen2.5-72B Open source Alibaba | 62.0% | — | 19 Sep 2024 | — | Source ↗ | |
| 2 | Llama 4 Maverick Open source Meta | 52.0% | — | 6 Apr 2025 | — | Source ↗ | |
| 3 | Gemma 3 27B Open source Google | 52.0% | — | 12 Mar 2025 | — | Source ↗ | |
| 4 | Llama 3.2 90B Open source Meta AI | 52.0% | — | 24 Sep 2024 | — | Source ↗ | |
| 5 | Pixtral 12B Open source Mistral AI | 37.0% | — | 17 Sep 2024 | — | Source ↗ |
Caveats: settings (reasoning effort, agent scaffold) differ between rows and materially affect scores; the best reported setting per model is shown. Source: Epoch AI, CC BY 4.0.