BAGEL - Wikipedia Cognition: leaderboard

BAGEL Wikipedia questions on animal cognition, a small dimension of 28 questions. Four-option closed-book questions written by GPT-4o-mini from source material that is withheld at test time; accuracy in percent, higher is better.

Metric: Accuracy (%, times 100) on the 28 Wikipedia cognition questions of BAGEL, closed-book four-option single-answer multiple choice (the source passage is withheld), one unified prompt, greedy decoding with seed 0, a single run per model; questions written by GPT-4o-mini from Wikipedia species articles, GloBI interaction records, bioRxiv preprints and Xeno-canto recordings, options shuffled; higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 13 models tracked. Updated 7 Oct 2026.

Top models

#ModelSettingScoreOverall rank
1Claude Opus 4.685.7#113
2Phi-485.7#1931
3Qwen 3 14BNon-reasoning82.1#1269
4GPT-5.478.6#135
5Gemma 3 27B (IT)75#1324
6Llama 3.1 8B Instruct64.3#4600
7Mistral 7B Instruct (v0.3)64.3#6750
8Qwen 3 8BNon-reasoning64.3#1873
9Qwen 3 32BNon-reasoning60.7#972
10Qwen 3 4BNon-reasoning60.7#2348
11gemma-7B57.1#7121
12SmolLM2-360M25#9570
13Qwen 3 0.6BNon-reasoning17.9#8350

No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.

Interactive version: theaggregate.ai/benchmark?slug=bagel-wikipedia-cognition · How It Works · Data refreshed daily, snapshot 2026-11-21.