APUN-Bench (Real-World) - Fine Pun Location: leaderboard
Metric: Fine pun-location accuracy (%), the predicted pun word exactly matching the annotated pun word after lemmatization, on the real-world set (893 human-recorded clips); temperature 0; higher is better. Source: arxiv.org. Saturation forecast: Estimated already saturated. 10 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 2.0 Flash | 51.03 | #331 |
| 2 | Qwen2.5-Omni-7B | 42.85 | #601 |
| 3 | MERaLiON-2-10B | 35.45 | #567 |
| 4 | Qwen2-Audio-7B-Instruct | 24.95 | #859 |
| 5 | SALMONN-13B | 14.11 | #1237 |
No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #8 Claude Fable 5.1.
Interactive version: theaggregate.ai/benchmark?slug=apun-bench-real-world-fine-pun-location · How It Works · Data refreshed daily, snapshot 2026-10-11.