AGC-Bench - simile_generation: leaderboard
The "simile_generation" 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: Not forecast. 81 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GLM-5.1 | 1.9 | #97 |
| 2 | GPT-4o | 1.84 | #343 |
| 3 | Gemma 4 26B A4B (IT) | 1.63 | #297 |
| 4 | ERNIE 4.5 300B A47B | 1.62 | #458 |
| 5 | GLM-4.7 | 1.35 | #192 |
| 6 | Gemma 3 27B (IT) | 1.34 | #523 |
| 7 | Gemma 4 31B (IT) | 1.32 | #176 |
| 8 | DeepSeek V3 (0324) | 0.94 | #340 |
| 9 | GPT-4.1 | 0.93 | #248 |
| 10 | Claude Sonnet 4.5 | 0.93 | #142 |
| 11 | pixtral-large-2411 | 0.92 | #478 |
| 12 | GPT-5.5 | 0.9 | #31 |
| 13 | GLM-5 | 0.87 | #141 |
| 14 | Claude Opus 4.6 | 0.82 | #63 |
| 15 | Claude Haiku 4.5 | 0.81 | #280 |
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-simile-generation · How It Works · Data refreshed daily, snapshot 2026-10-20.