VisReason (Vision-Centric) - Localized Reasoning: leaderboard
VisReason Localized Reasoning is a perceptual-reasoning category of VisReason (2026): the model must infer the target instance in an image and localize it with a bounding box. The score is accuracy in percent, higher is better.
Metric: Accuracy (%) on the Localized Reasoning category (perceptual reasoning: infer the target instance in an image and localize it with a bounding box) of VisReason, answers scored by format (regular-expression match, GPT-5-mini judge for open-ended answers, IoU above 0.5 for boxes), zero-shot with chain-of-thought prompt templates, one per answer format; rows not shaded gray in the paper use explicit reasoning (GPT-5 and Gemini 3 Pro at low reasoning effort); higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 20 models tracked. Updated 7 Oct 2026.
Top models
| # | Model | Setting | Score | Overall rank |
|---|---|---|---|---|
| 1 | Gemini 3 Pro (Preview) | Low | 44.2 | #123 |
| 2 | GPT-5.2 | Thinking | 24.6 | #190 |
| 3 | GPT-5 | Low | 15 | #185 |
| 4 | GPT-5 Mini | High | 12.5 | #373 |
| 5 | Qwen 3 VL 8B (Thinking) | 9.6 | #924 | |
| 6 | Qwen 3 VL 235B A22B (Thinking) | 7.1 | #462 | |
| 7 | GPT-4o | 6.3 | #754 | |
| 8 | GPT-5 Nano | 3.8 | #912 | |
| 9 | O4 Mini | 3.8 | #372 | |
| 10 | Qwen 2.5 VL 32B Instruct | 3.3 | #1065 | |
| 11 | Qwen 3 VL 2B (Thinking) | 2.5 | #2552 | |
| 12 | MiMo-VL-7B-RL | 2.1 | #1160 | |
| 13 | Qwen 3 VL 8B Instruct | 1.3 | #1043 | |
| 14 | InternVL3-14B | 1.3 | #1227 | |
| 15 | Qwen 3 VL 32B (Thinking) | 1.3 | #577 |
No result here: #2 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.
Interactive version: theaggregate.ai/benchmark?slug=visreason-vision-centric-localized-reasoning · How It Works · Data refreshed daily, snapshot 2026-11-22.