AGC-Bench - outline_to_story: leaderboard
The "outline_to_story" 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 August 2027. 82 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemma 4 31B (IT) | 1.81 | #176 |
| 2 | GLM-5 | 1.74 | #141 |
| 3 | GLM-5.1 | 1.58 | #97 |
| 4 | GLM-5-Turbo | 1.45 | #119 |
| 5 | Gemma 4 26B A4B (IT) | 1.34 | #297 |
| 6 | GPT-5.5 | 1.31 | #31 |
| 7 | GLM-4.7 | 1.25 | #192 |
| 8 | DeepSeek V3.1 Terminus | 1.14 | #220 |
| 9 | Llama 3.3 70B Instruct | 1.11 | #534 |
| 10 | Llama 4 Scout | 1.11 | #656 |
| 11 | GPT-5.4 | 1.06 | #79 |
| 12 | Llama 3.1 70B Instruct | 1.05 | #561 |
| 13 | DeepSeek V3.2 Exp | 1.04 | #234 |
| 14 | DeepSeek V3.2 | 1.03 | #205 |
| 15 | jamba-large-1.7 | 1.02 | #525 |
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-outline-to-story · How It Works · Data refreshed daily, snapshot 2026-10-20.