AGC-Bench - mops: leaderboard
The "mops" 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 | Mistral Medium 3.1 | 2.56 | #408 |
| 2 | Llama 3 8B Instruct | 2.05 | #1129 |
| 3 | Ministral 3 8B | 1.99 | #687 |
| 4 | Claude Sonnet 4.6 | 1.68 | #90 |
| 5 | Nemotron 3 Super | 1.53 | #335 |
| 6 | Mistral Large 3 | 1.46 | #401 |
| 7 | Llama 3.1 70B Instruct | 1.32 | #561 |
| 8 | ERNIE-4.5-VL-424B-A47B | 1.26 | #391 |
| 9 | Gemma 3 4B (IT) | 1.08 | #992 |
| 10 | Nova 2.0 Lite (Non-reasoning) | 1.05 | #417 (Nova 2.0 Lite) |
| 11 | Claude Opus 4.6 | 1.02 | #63 |
| 12 | DeepSeek V3 (0324) | 1.01 | #340 |
| 13 | Claude Opus 4.7 | 1 | #49 |
| 14 | Kimi K2 0905 | 1 | #288 |
| 15 | GPT-5.1 | 0.96 | #135 |
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-mops · How It Works · Data refreshed daily, snapshot 2026-10-20.