AGC-Bench - munch: leaderboard
The "munch" 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: Estimated already saturated. 83 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-4.1 | 1.18 | #248 |
| 2 | GPT-5.5 | 1.18 | #31 |
| 3 | GPT-4.1 Mini | 1.06 | #355 |
| 4 | GPT-5.4 | 1.04 | #79 |
| 5 | Claude Sonnet 4.6 | 0.98 | #90 |
| 6 | Claude Opus 4.7 | 0.98 | #49 |
| 7 | Gemma 4 31B (IT) | 0.98 | #176 |
| 8 | GPT-5.1 | 0.98 | #135 |
| 9 | GLM-5.1 | 0.98 | #97 |
| 10 | Claude Opus 4.5 | 0.98 | #83 |
| 11 | ERNIE 4.5 300B A47B | 0.98 | #458 |
| 12 | Qwen Turbo | 0.98 | #699 |
| 13 | GPT-5.4 Mini | 0.94 | #212 |
| 14 | DeepSeek V3.2 | 0.9 | #205 |
| 15 | GLM-4.7 | 0.9 | #192 |
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-munch · How It Works · Data refreshed daily, snapshot 2026-10-20.