CommunityBench - Community-Consistent Generation: leaderboard
Open-ended replies written in a given Reddit community's voice, compared pairwise by GPT-4o, Grok-4-Fast and Gemini-2.5-Flash with majority voting and fitted to a Bradley-Terry Elo rating; GPT-4o is also one of the models judged. The rating is unbounded and relative to the models compared together; higher is better (CommunityBench, Jiayu Lin and Zhongyu Wei, Fudan University, 2026).
Metric: BTL-Elo rating from majority-voted LLM-judge pairwise comparisons. Source: arxiv.org. Saturation forecast: Estimated already saturated. 17 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | DeepSeek R1 0528 | 812.3 | #425 |
| 2 | Grok 4 | 478.67 | #306 |
| 3 | Qwen 3 14B | 272.39 | #1235 |
| 4 | GPT-4o | 256.28 | #741 |
| 5 | Qwen 3 32B | 237.8 | #960 |
| 6 | GLM-4 32B (0414) | 233.49 | #1407 |
| 7 | DeepSeek V3 (0324) | 206.22 | #704 |
| 8 | glm-4-9B (0414) | 176.85 | #2196 |
| 9 | Qwen 3 8B | 87.77 | #1838 |
| 10 | Llama 3.1 70B Instruct | -54.9 | #1541 |
| 11 | Llama 3.3 70B Instruct | -56.64 | #1353 |
| 12 | Qwen 2.5 72B Instruct | -149.18 | #1093 |
| 13 | Qwen 2.5 14B Instruct | -154.87 | #1745 |
| 14 | Llama 3.1 8B Instruct | -166.28 | #4277 |
| 15 | Qwen 2.5 7B Instruct | -262.68 | #2583 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=communitybench-community-consistent-generation · How It Works · Data refreshed daily, snapshot 2026-11-12.