NLU Corruption - Clean RTE: leaderboard

This study tests natural-language inference with clean and corrupted inputs. Raw accuracy, answer-bias-centred accuracy and prompting conditions remain separate. The public ANLI-trained RoBERTa checkpoint is distinct from generic RoBERTa and the study’s locally fine-tuned encoders. Above-chance retention is not relabelled as accuracy. This board reports clean answer accuracy (%). Higher is better.

Metric: Clean answer accuracy (%). Source: arxiv.org. Saturation forecast: Estimated already saturated. 8 models tracked. Updated 14 Nov 2026.

Top models

#ModelSettingScoreOverall rank
1Qwen 3.8 27B91.3#209
2OLMo 3 32B Base90.6
3Grok 4.7Reasoning88.8#47
4DeepSeek V4 Pro88.5#191
5OLMo 3 7B Base87
6Mistral Large 385.6#827
7Muse Glimmer 30B80.9#391
8GPT-4o80.1#744

No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.

Interactive version: theaggregate.ai/benchmark?slug=nlu-corruption-clean-rte · How It Works · Data refreshed daily, snapshot 2026-11-16.