MMOOC - Unclear Logical and Symbolic (Yes/No): leaderboard

Metric: Out-of-context score (%; mean of the refusal rate, the share of the Unclear Logical and Symbolic Yes/No questions, which the image and context cannot answer, that the model correctly declines, and refusal rationality, the mean 0-1 judge score of its abstention reasoning in steps of 0.25; judge scores averaged over GPT-5.6, Claude Opus 5 and DeepSeek-V4-Pro). Source: arxiv.org. Saturation forecast: Around 2033. 18 models tracked.

Top models

#ModelScore
1O154.75
2Gemma 4 26B49
3Gemma 4 31B46.75
4Ministral 3 14B45.75
5InternVL3-8B39.25
6O334.25
7GPT-4o33.75
8Ministral 3 8B31
9Claude Opus 4.630.5
10Qwen 3.5 122B A10B28.25
11Qwen 3.5 27B28
12Gemini 3.1 Pro (Preview)27.75
13Llama 4 Maverick27.75

Interactive version: theaggregate.ai/benchmark?slug=mmooc-unclear-logical-and-symbolic-yes-no · How It Works · Data refreshed daily, snapshot 2026-09-29.