GLUE MNLI (Mismatched): leaderboard
Multi-Genre NLI (Williams et al., 2018): label sentence pairs as entailment, neutral or contradiction. This board is accuracy on the mismatched set, whose genres are absent from the training data, from the archived Papers With Code table of scores as each paper reported them; higher is better.
Metric: Mismatched. Source: huggingface.co. Saturation forecast: Rough model projection: around 2026. 45 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | T5 (11B) | 91.7 | #1631 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=glue-mnli-mismatched · How It Works · Data refreshed daily, snapshot 2026-11-05.