LOL Bench - Joke Explanation: leaderboard
Blind-graded benchmark of how well language models explain jokes, scored as a mean grade over a fixed item pool with three samples per item.
Metric: Mean Grade (%). Source: www.lolbench.lol. Saturation forecast: Estimated already saturated. 13 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 3.8 Max | 91.67 | #34 |
| 2 | Grok 4.6 | 90.65 | #29 |
| 3 | Qwen 3.8 Flash | 90.56 | #62 |
| 4 | GLM-5.3 Flash | 90.45 | #77 |
| 5 | GLM-5.3 | 90.38 | #44 |
| 6 | Claude Opus 5 | 90.02 | #16 |
| 7 | GPT-5.6 Pro Sol | 89.11 | #21 |
| 8 | Muse Spark 1.2 | 88.22 | #69 |
| 9 | Gemini 3.1 Pro (Preview) | 88.15 | #64 |
| 10 | DeepSeek V4 Pro | 87.26 | #112 |
| 11 | Hy3 | 86.74 | #100 |
| 12 | MiMo-V2.5-Pro | 83.73 | #167 |
| 13 | MiMo-V2.5 | 83.65 | #196 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=lol-bench-joke-explanation · How It Works · Data refreshed daily, snapshot 2026-11-05.