AGC-Bench - sdat: leaderboard
The "sdat" 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. 82 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 4.6 | 0.68 | #63 |
| 2 | Llama 4 Scout | 0.67 | #656 |
| 3 | Gemma 3 4B (IT) | 0.67 | #992 |
| 4 | Claude Opus 4.5 | 0.62 | #83 |
| 5 | Llama 3.3 70B Instruct | 0.62 | #534 |
| 6 | Claude Sonnet 4 | 0.61 | #199 |
| 7 | Hermes-3-Llama-3.1-405B | 0.61 | #510 |
| 8 | Claude 3.7 Sonnet | 0.58 | #249 |
| 9 | Llama 3.1 70B Instruct | 0.57 | #561 |
| 10 | Claude Sonnet 4.5 | 0.56 | #142 |
| 11 | Llama 3.2 11B Instruct | 0.56 | #1126 |
| 12 | Llama 3.2 3B Instruct | 0.54 | #1352 |
| 13 | Gemma 2 27B (IT) | 0.53 | #645 |
| 14 | Claude Opus 4.7 | 0.51 | #49 |
| 15 | Llama 3 8B Instruct | 0.51 | #1129 |
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-sdat · How It Works · Data refreshed daily, snapshot 2026-10-20.