AGC-Bench - story_generation_rocstories: leaderboard
The "story_generation_rocstories" 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. 83 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Olmo 3.1 32B Instruct | 3.29 | #763 |
| 2 | jamba-large-1.7 | 2.32 | #525 |
| 3 | GPT-4.1 Mini | 2.26 | #355 |
| 4 | DeepSeek V3 (0324) | 1.88 | #340 |
| 5 | GPT-4.1 | 1.34 | #248 |
| 6 | DeepSeek V3 Chat | 1.32 | #447 |
| 7 | Claude Sonnet 4.5 | 1.29 | #142 |
| 8 | minimax-m2-her | 1.21 | |
| 9 | Llama 3 8B Instruct | 1.06 | #1129 |
| 10 | Mistral Small 3.1 | 1.04 | #610 |
| 11 | trinity-large-preview | 0.89 | #582 |
| 12 | Hermes-3-Llama-3.1-405B | 0.89 | #510 |
| 13 | Llama 4 Maverick | 0.79 | #464 |
| 14 | Llama 3.1 8B Instruct | 0.77 | #1035 |
| 15 | NVIDIA Nemotron Nano 9B V2 | 0.77 | #823 |
Interactive version: theaggregate.ai/benchmark?slug=agc-bench-story-generation-rocstories · How It Works · Data refreshed daily, snapshot 2026-10-20.