AGC-Bench - moh_x: leaderboard
The "moh_x" 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 April 2027. 82 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemma 4 31B (IT) | 0.88 | #176 |
| 2 | GLM-5 | 0.88 | #141 |
| 3 | GPT-4.1 | 0.87 | #248 |
| 4 | Claude Sonnet 4.6 | 0.82 | #90 |
| 5 | Claude Opus 4.6 | 0.82 | #63 |
| 6 | Claude Sonnet 4 | 0.82 | #199 |
| 7 | Claude 3.7 Sonnet | 0.82 | #249 |
| 8 | GPT-5.1 | 0.82 | #135 |
| 9 | Claude Opus 4.5 | 0.82 | #83 |
| 10 | minimax-01 | 0.82 | #606 |
| 11 | GPT-5.4 | 0.79 | #79 |
| 12 | Gemini 2.5 Flash | 0.76 | #245 |
| 13 | Claude Haiku 4.5 | 0.76 | #280 |
| 14 | ERNIE 4.5 300B A47B | 0.76 | #458 |
| 15 | Hermes 4 405B | 0.76 | #334 |
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-moh-x · How It Works · Data refreshed daily, snapshot 2026-10-20.