Logical-CommonsenseQA - AND: leaderboard

Metric: Macro-F1 (%; both composed statements must be plausible; zero-shot prompt, or the authors' fine-tunes on the training split; 250 human-validated four-option test questions). Source: arxiv.org. Saturation forecast: Around December 2026. 8 models tracked.

Top models

#ModelScore
1Gemini 2.5 Flash84.1
2Gemini 3 Flash (Preview)82.5
3Llama 3.3 70B Instruct80.9
4Qwen 2.5 7B Instruct79.6
5Llama 3.1 8B Instruct71.9

Interactive version: theaggregate.ai/benchmark?slug=logical-commonsenseqa-and · How It Works · Data refreshed daily, snapshot 2026-09-26.