AGC-Bench - scimon: leaderboard
The "scimon" 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. 82 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | ERNIE 4.5 300B A47B | 4.03 | #458 |
| 2 | Gemma 4 31B (IT) | 1.88 | #176 |
| 3 | GLM-5-Turbo | 1.79 | #119 |
| 4 | Mistral Small 3.1 | 1.5 | #610 |
| 5 | Claude Opus 4.7 | 1.37 | #49 |
| 6 | GLM-5 | 1.34 | #141 |
| 7 | Gemini 2.5 Flash | 1.27 | #245 |
| 8 | Claude Opus 4.5 | 1.1 | #83 |
| 9 | Claude 3.7 Sonnet | 1.08 | #249 |
| 10 | Gemma 4 26B A4B (IT) | 1.05 | #297 |
| 11 | GLM-5.1 | 0.7 | #97 |
| 12 | GPT-4.1 | 0.69 | #248 |
| 13 | Llama 4 Scout | 0.69 | #656 |
| 14 | Nova Micro (v1) | 0.6 | #831 |
| 15 | Kimi K2.6 | 0.55 | #104 |
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-scimon · How It Works · Data refreshed daily, snapshot 2026-10-20.