ASCIIArt-Bench - Generation (Recall): leaderboard
Metric: Weighted composite score (x100) of five Gemini 3 Pro judge dimensions scored 0 to 1 from the rendered image and raw text (instruction faithfulness 0.35, semantic alignment 0.25, structural coherence 0.15, spatial logic 0.15, character efficiency 0.10), text-to-ASCII-art generation on ASCIIArt-Bench's Recall subset (instructions in the distribution of the ASCIIArt-7K training data); higher is better. Source: arxiv.org. Saturation forecast: Around March 2027. 10 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Sonnet 4.5 | 74.2 | #138 |
| 2 | GPT-5.2 | 63.3 | #105 |
| 3 | Kimi K2 | 53.7 | #236 |
| 4 | Gemini 3 Flash | 52.2 | #93 |
| 5 | Qwen 2.5 72B Instruct | 48.5 | #436 |
| 6 | Qwen 2.5 32B Instruct | 47.5 | #491 |
| 7 | DeepSeek V3.2 | 47.4 | #198 |
| 8 | Seed-1.6 | 41.3 | #257 |
| 9 | Qwen 2.5 7B Instruct | 39.8 | #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-generation-recall · How It Works · Data refreshed daily, snapshot 2026-10-11.