HAERAE-Vision - Business/Economics: leaderboard

The 37 business and economics questions of HAERAE-Vision (What Users Leave Unsaid, Choi et al., ACL 2026 Findings): real Korean community questions on economics, accounting and management, asked with images in the users' original under-specified wording. A GPT-5-mini judge scores each answer against a 1-5 item checklist (met 1, partly met 0.5, not met 0); the score is the mean checklist credit per question, in percent, averaged over three runs; higher is better.

Metric: Checklist Score (%; original under-specified queries, GPT-5-mini judge, mean of 3 runs). Source: arxiv.org. Saturation forecast: Around February 2027. 45 models tracked.

Top models

#ModelScoreOverall rank
1GPT-558.41#163
2Gemini 2.5 Pro51.09#262
3GPT-5 Mini50.19#349
4Gemini 2.5 Flash48.05#452
5Qwen 3 VL 235B A22B (Thinking)47.3#451
6Grok 444.44#306
7Qwen 3 VL 235B A22B Instruct43.16#521
8Qwen 3 VL 30B A3B Instruct42.02#870
9Qwen 3 VL 32B Instruct41.29#582
10Gemini 2.5 Flash Lite38.84#958
11Qwen 3 VL 30B A3B (Thinking)38.44#710
12Qwen 3 VL 32B (Thinking)38.21#557
13Qwen 3 VL 8B (Thinking)33.92#887
14Gemma 3 27B31.71#1638
15Qwen 3 VL 4B Instruct31.04#1281

No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.

Interactive version: theaggregate.ai/benchmark?slug=haerae-vision-business-economics · How It Works · Data refreshed daily, snapshot 2026-11-12.