AGC-Bench - fann_or_flop: leaderboard
The "fann_or_flop" 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. 83 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | trinity-large-preview | 2.3 | #582 |
| 2 | Llama 4 Maverick | 1.74 | #464 |
| 3 | GLM-5.1 | 1.37 | #97 |
| 4 | ERNIE 4.5 300B A47B | 1.36 | #458 |
| 5 | GPT-5.1 | 1.27 | #135 |
| 6 | Kimi K2 0905 | 1.16 | #288 |
| 7 | GLM-4.7 | 1.14 | #192 |
| 8 | Claude 3.7 Sonnet | 1.13 | #249 |
| 9 | GPT-4o | 1.04 | #343 |
| 10 | Claude Sonnet 4.5 | 1.02 | #142 |
| 11 | ERNIE-4.5-VL-424B-A47B | 0.95 | #391 |
| 12 | DeepSeek V3 Chat | 0.92 | #447 |
| 13 | GLM-5-Turbo | 0.83 | #119 |
| 14 | DeepSeek V3 (0324) | 0.77 | #340 |
| 15 | Hermes 4 405B | 0.74 | #334 |
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-fann-or-flop · How It Works · Data refreshed daily, snapshot 2026-10-20.