AGC-Bench - metaphoric_analogies: leaderboard
The "metaphoric_analogies" 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. 83 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 4.7 | 3.98 | #49 |
| 2 | Gemma 4 31B (IT) | 3.21 | #176 |
| 3 | GLM-5 | 2.69 | #141 |
| 4 | Gemini 2.5 Flash | 2.36 | #245 |
| 5 | Hermes 4 405B | 2.18 | #334 |
| 6 | GLM-5.1 | 1.42 | #97 |
| 7 | GPT-5.5 | 1.39 | #31 |
| 8 | Gemma 4 26B A4B (IT) | 1.27 | #297 |
| 9 | GLM-4.6 | 1.16 | #252 |
| 10 | GPT-5.4 | 1.15 | #79 |
| 11 | GPT-5.1 | 0.89 | #135 |
| 12 | Kimi K2.5 | 0.84 | #144 |
| 13 | Olmo 3.1 32B Instruct | 0.83 | #763 |
| 14 | GPT-5.4 Mini | 0.78 | #212 |
| 15 | Hunyuan A13B-Instruct | 0.63 | #666 |
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-metaphoric-analogies · How It Works · Data refreshed daily, snapshot 2026-10-20.