AGC-Bench - hypogen: leaderboard
The "hypogen" 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. Saturation forecast: Around June 2027. 83 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Kimi K2 0905 | 2.19 | #288 |
| 2 | Kimi K2.5 | 2.04 | #144 |
| 3 | Kimi K2.6 | 1.95 | #104 |
| 4 | GPT-5.1 | 1.85 | #135 |
| 5 | GPT-5.4 | 1.53 | #79 |
| 6 | DeepSeek V3.2 Exp | 1.51 | #234 |
| 7 | Claude Sonnet 4.5 | 1.5 | #142 |
| 8 | Qwen 3 Next 80B A3B Instruct | 1.44 | #368 |
| 9 | Mistral Medium 3.1 | 1.28 | #408 |
| 10 | GLM-5.1 | 1.24 | #97 |
| 11 | DeepSeek V3.1 Terminus | 1.23 | #220 |
| 12 | DeepSeek V3.2 | 1.17 | #205 |
| 13 | Claude Sonnet 4.6 | 1 | #90 |
| 14 | Claude Sonnet 4 | 0.97 | #199 |
| 15 | DeepSeek V3.1 NEX N1 | 0.94 | #327 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=agc-bench-hypogen · How It Works · Data refreshed daily, snapshot 2026-10-20.