HAERAE-Vision - Shopping/Consumer: leaderboard
Metric: Checklist Score (%; original under-specified queries, GPT-5-mini judge, mean of 3 runs). Source: arxiv.org. Saturation forecast: Around December 2026. 45 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5 | 55.96 |
| 2 | Gemini 2.5 Flash | 46.13 |
| 3 | Gemini 2.5 Pro | 46 |
| 4 | GPT-5 Mini | 44.33 |
| 5 | Qwen 3 VL 32B Instruct | 43.38 |
| 6 | Qwen 3 VL 235B A22B Instruct | 40.98 |
| 7 | Qwen 3 VL 32B (Thinking) | 38.16 |
| 8 | Qwen 3 VL 30B A3B (Thinking) | 37.99 |
| 9 | Grok 4 | 36.73 |
| 10 | Qwen 3 VL 30B A3B Instruct | 36.53 |
| 11 | Qwen 3 VL 235B A22B (Thinking) | 34.18 |
| 12 | Qwen 3 VL 8B (Thinking) | 32.53 |
| 13 | Qwen 3 VL 4B (Thinking) | 31.29 |
| 14 | Gemini 2.5 Flash Lite | 28.82 |
| 15 | Qwen 3 VL 8B Instruct | 27.98 |
Interactive version: theaggregate.ai/benchmark?slug=haerae-vision-shopping-consumer · How It Works · Data refreshed daily, snapshot 2026-09-25.