AGC-Bench - story_generation_rocstories — leaderboard
Metric: Dataset z-score. Source: huggingface.co. 83 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Olmo 3.1 32B Instruct | 3.29 |
| 2 | jamba-large-1.7 | 2.32 |
| 3 | GPT-4.1 Mini | 2.26 |
| 4 | DeepSeek V3 (0324) | 1.88 |
| 5 | GPT-4.1 | 1.34 |
| 6 | DeepSeek V3 Chat | 1.32 |
| 7 | Claude Sonnet 4.5 | 1.29 |
| 8 | Llama 3 8B Instruct | 1.06 |
| 9 | Mistral Small 3.1 | 1.04 |
| 10 | trinity-large-preview | 0.89 |
| 11 | Llama 4 Maverick | 0.79 |
| 12 | Llama 3.1 8B Instruct | 0.77 |
| 13 | NVIDIA Nemotron Nano 9B V2 | 0.77 |
| 14 | Llama 3.1 70B Instruct | 0.67 |
| 15 | Hermes 4 405B | 0.64 |
Interactive version: theaggregate.ai/benchmark?slug=agc-bench-story-generation-rocstories · How the rankings work · Data refreshed daily, snapshot 2026-07-22.