YOMI-Bench - Rhyme Generation: leaderboard
Generate a hiragana word whose vowel sequence exactly matches a given hiragana word, checked with a fixed vowel conversion table; from YOMI-Bench (2026). Accuracy in percent averaged over five paraphrased prompts; higher is better.
Metric: Accuracy (%; generated hiragana word whose vowel sequence exactly matches the input word under a fixed vowel conversion table; 120 items; mean over five paraphrased prompts; accuracy as a fraction x100). Source: arxiv.org. Saturation forecast: Around July 2027. 10 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Llama-3.1-Swallow-8B-Instruct-v0.5 | 84.8 | #2704 |
| 2 | GPT-5 | 78 | #167 |
| 3 | Claude Sonnet 4.5 | 72.8 | #250 |
| 4 | Ministral-8B-Instruct-2410 | 65.19 | #4684 |
| 5 | GPT-4o | 59.8 | #741 |
| 6 | Gemini 2.5 Flash | 59.2 | #453 |
| 7 | Mistral Medium 3.1 | 46.79 | #852 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=yomi-bench-rhyme-generation · How It Works · Data refreshed daily, snapshot 2026-11-13.