GST-Bench - Top-Down Selection (Hard): leaderboard

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 December 2027. 22 models tracked.

Top models

#ModelScore
1Gemini 3 Pro52.25
2GPT-4o45.51
3GPT-543.26
4Gemini 2.5 Pro41.01
5Seed 1.837.64
6Qwen 3 VL 32B32.58
7InternVL3.5-8B25.84

Interactive version: theaggregate.ai/benchmark?slug=gst-bench-top-down-selection-hard · How It Works · Data refreshed daily, snapshot 2026-09-29.