GST-Bench - Top-Down Selection (Hard): leaderboard
GST-Bench (2026) asks vision-language models about long egocentric videos rendered in a simulator, where answers require a global map of the scene rather than the current frame. The model picks the top-down image that matches the explored scene among hard distractors; accuracy in percent; higher is better.
Metric: Accuracy (%; multiple choice of the top-down image of the explored scene, hard distractors; zero-shot, greedy decoding, official prompt templates, over human-verified questions on 6,790 minutes of simulator-rendered egocentric exploration video). Source: arxiv.org. Saturation forecast: Around January 2028. 22 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3 Pro | 52.25 | #137 |
| 2 | GPT-4o | 45.51 | #741 |
| 3 | GPT-5 | 43.26 | #167 |
| 4 | Gemini 2.5 Pro | 41.01 | #266 |
| 5 | Seed 1.8 | 37.64 | #242 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=gst-bench-top-down-selection-hard · How It Works · Data refreshed daily, snapshot 2026-11-13.