ESPIRE - Pick Localization: leaderboard
Metric: Localization accuracy (%) on the 1,095 pick tasks: the predicted 2D target point (with ground-truth depth, at most 3 attempts) lands on a correct target; no reflection input; higher is better. Source: arxiv.org. Saturation forecast: Around January 2027. 6 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 2.5 Pro | 57.72 | #145 |
| 2 | Qwen 3 VL 30B A3B Instruct | 54.43 | #365 |
| 3 | Qwen 3 VL 235B A22B Instruct | 51.96 | #264 |
| 4 | Qwen 3 VL 8B Instruct | 47.03 | #401 |
| 5 | InternVL3-78B | 28.31 | #345 |
No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #8 Claude Fable 5.1.
Interactive version: theaggregate.ai/benchmark?slug=espire-pick-localization · How It Works · Data refreshed daily, snapshot 2026-10-11.