ASCIIArt-Bench - Category Selection: leaderboard
Metric: Accuracy (%) on ASCIIArt-Bench's 200 ASCII-art understanding items, mean of the seen-category and unseen-category halves; a Gemini 3 Pro judge checks whether the category picked from a candidate list matches the label; higher is better. Source: arxiv.org. Saturation forecast: Around January 2027. 10 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3 Flash | 38.5 | #93 |
| 2 | Claude Sonnet 4.5 | 20 | #138 |
| 3 | GPT-5.2 | 17.5 | #105 |
| 4 | Qwen 2.5 72B Instruct | 7.5 | #436 |
| 5 | Kimi K2 | 7.5 | #236 |
| 6 | DeepSeek V3.2 | 7 | #198 |
| 7 | Qwen 2.5 32B Instruct | 7 | #491 |
| 8 | Seed-1.6 | 4.5 | #257 |
| 9 | Qwen 2.5 7B Instruct | 2.5 | #846 |
No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #8 Claude Fable 5.1.
Interactive version: theaggregate.ai/benchmark?slug=asciiart-bench-category-selection · How It Works · Data refreshed daily, snapshot 2026-10-11.