NICE - Self-Consistency: leaderboard
Metric: Ranking accuracy (%) on the self-consistency dimension on NICE (137 human-written items in Chinese social contexts, each a ranking of an optimal, a suboptimal and a boundary-violating response), an item counts only when the full ranking matches the gold order, first-party APIs at temperature 0, three runs pooled; higher is better. Source: arxiv.org. Saturation forecast: Around January 2027. 5 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5.5 | 80.6 |
| 2 | DeepSeek V4 Pro | 80.6 |
| 3 | Gemini 3.1 Pro (Preview) | 66.7 |
| 4 | Claude Opus 4.7 | 66.7 |
| 5 | Qwen 3.6 Plus | 66.7 |
Interactive version: theaggregate.ai/benchmark?slug=nice-self-consistency · How It Works · Data refreshed daily, snapshot 2026-10-07.