AGC-Bench - unfun_corpus: leaderboard
The "unfun_corpus" subtask of AGC-Bench. A task from AGC-Bench, evaluating models on open-ended creative and figurative-language generation — humor, metaphor, analogy, storytelling and brainstorming — typically graded by LLM-as-judge.
Metric: Dataset z-score. Source: huggingface.co. 79 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | nemotron-3-nano-30B-a3B | 4.91 | #662 |
| 2 | Qwen 3 14B | 1.44 | #537 |
| 3 | Hermes 3 - Llama-3.1 70B | 1.37 | #927 |
| 4 | Mistral Medium 3.1 | 1.03 | #408 |
| 5 | minimax-m2-her | 0.75 | |
| 6 | Llama 4 Scout | 0.73 | #656 |
| 7 | Ministral 3 8B | 0.72 | #687 |
| 8 | Gemma 3 27B (IT) | 0.66 | #523 |
| 9 | ERNIE-4.5-VL-424B-A47B | 0.63 | #391 |
| 10 | Qwen 3 Next 80B A3B Instruct | 0.59 | #368 |
| 11 | Kimi K2 0905 | 0.56 | #288 |
| 12 | Gemma 4 26B A4B (IT) | 0.53 | #297 |
| 13 | DeepSeek V3 Chat | 0.38 | #447 |
| 14 | Gemma 4 31B (IT) | 0.36 | #176 |
| 15 | Claude Haiku 4.5 | 0.35 | #280 |
Interactive version: theaggregate.ai/benchmark?slug=agc-bench-unfun-corpus · How It Works · Data refreshed daily, snapshot 2026-10-20.