AGC-Bench - slang_generation: leaderboard
The "slang_generation" 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 July 2027. 83 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Sonnet 4.5 | 1.52 | #142 |
| 2 | DeepSeek V3.2 | 1.26 | #205 |
| 3 | Claude Opus 4.5 | 1.24 | #83 |
| 4 | GLM-5 | 1.21 | #141 |
| 5 | Claude Sonnet 4 | 1.14 | #199 |
| 6 | Claude Opus 4.6 | 1.13 | #63 |
| 7 | Llama 3.2 11B Instruct | 1.1 | #1126 |
| 8 | Gemma 4 31B (IT) | 1.03 | #176 |
| 9 | Gemini 2.5 Flash | 1.03 | #245 |
| 10 | GLM-5-Turbo | 1 | #119 |
| 11 | Gemma 4 26B A4B (IT) | 0.99 | #297 |
| 12 | DeepSeek V3.1 | 0.97 | #264 |
| 13 | DeepSeek V3.1 NEX N1 | 0.92 | #327 |
| 14 | DeepSeek V3.2 Exp | 0.91 | #234 |
| 15 | GPT-5.4 Mini | 0.81 | #212 |
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-slang-generation · How It Works · Data refreshed daily, snapshot 2026-10-20.