Paraphrase-Induced Output-Mode Collapse — leaderboard
Metric: Semantic Consistency Score (self-reported). Source: benchmarklist.com. 5 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Sonnet 4.5 | 41.1 |
| 2 | Claude Haiku 4.5 | 34.4 |
| 3 | GPT-4o Mini | 28.5 |
| 4 | GPT-4.1 Mini | 26.2 |
| 5 | Gemini 2.5 Flash | 25.6 |
Interactive version: theaggregate.ai/benchmark?slug=paraphrase-induced-output-mode-collapse · How the rankings work · Data refreshed daily, snapshot 2026-07-22.