AGC-Bench - ttcw: leaderboard
The "ttcw" 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 April 2027. 83 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 4.6 | 2.27 | #63 |
| 2 | GLM-5-Turbo | 2.27 | #119 |
| 3 | Claude Opus 4.7 | 1.83 | #49 |
| 4 | Qwen 3.5 Flash (02-23) | 1.77 | #185 |
| 5 | Nemotron 3 Super | 1.68 | #335 |
| 6 | Claude Sonnet 4.5 | 1.65 | #142 |
| 7 | GPT-5.4 | 1.44 | #79 |
| 8 | GLM-5.1 | 1.33 | #97 |
| 9 | Claude Opus 4.5 | 1.33 | #83 |
| 10 | Gemma 4 31B (IT) | 1.3 | #176 |
| 11 | Claude Haiku 4.5 | 1.24 | #280 |
| 12 | Gemini 2.5 Flash | 1.12 | #245 |
| 13 | Claude Sonnet 4.6 | 1.06 | #90 |
| 14 | GPT-5.1 | 0.89 | #135 |
| 15 | DeepSeek V3.1 Terminus | 0.85 | #220 |
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-ttcw · How It Works · Data refreshed daily, snapshot 2026-10-20.