MEGA-Bench Task - VLN Identify Location: leaderboard
Metric: Task Score (%). Source: huggingface.co. 44 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | InternVL3-78B | 57.6 |
| 2 | Gemini 2.5 Pro (Preview 03-25) | 53.3 |
| 3 | Gemini 1.5 Pro (002) | 39.4 |
| 4 | Claude 3.5 Sonnet (20240620) | 38.8 |
| 5 | GPT-4o | 32.7 |
| 6 | InternVL3-8B | 27.3 |
| 7 | Gemini 1.5 Flash (002) | 23 |
| 8 | Claude 3.5 Sonnet (20241022) | 21.8 |
| 9 | Gemma 3 27B (IT) | 21.2 |
| 10 | Gemma 3 12B (IT) | 21.2 |
| 11 | Gemma 3 4B (IT) | 21.2 |
| 12 | InternVL3-38B | 20.6 |
| 13 | InternVL2.5-78B | 15.2 |
| 14 | Pixtral-12B | 14.5 |
| 15 | Qwen 2 VL 72B | 13.3 |
Interactive version: theaggregate.ai/benchmark?slug=mega-bench-task-vln-identify-location · How It Works · Data refreshed daily, snapshot 2026-09-05.