AGC-Bench - tinystories: leaderboard
The "tinystories" 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 | Qwen 3 Next 80B A3B Instruct | 1.8 | #368 |
| 2 | Kimi K2 0905 | 1.67 | #288 |
| 3 | DeepSeek V3.2 Exp | 1.64 | #234 |
| 4 | Mistral Medium 3.1 | 1.6 | #408 |
| 5 | Ministral 3 8B | 1.56 | #687 |
| 6 | ERNIE-4.5-VL-424B-A47B | 1.37 | #391 |
| 7 | DeepSeek V3.1 Terminus | 1.23 | #220 |
| 8 | Gemini 2.5 Flash | 1.2 | #245 |
| 9 | GPT-5.1 | 1.19 | #135 |
| 10 | Gemma 3 27B (IT) | 1.09 | #523 |
| 11 | Kimi K2.6 | 1.02 | #104 |
| 12 | Mixtral 8x22B Instruct | 0.94 | #728 |
| 13 | GPT-5.4 Nano | 0.91 | #328 |
| 14 | DeepSeek V3.1 | 0.88 | #264 |
| 15 | nemotron-3-nano-30B-a3B | 0.81 | #662 |
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-tinystories · How It Works · Data refreshed daily, snapshot 2026-10-20.