BAGEL - Wikipedia Geographic Distribution: leaderboard

Metric: Accuracy (%, times 100) on the 281 Wikipedia geographic distribution 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: Estimated already saturated. 13 models tracked.

Top models

#ModelScore
1Claude Opus 4.698.9
2GPT-5.497.9
3Qwen 3 32B (Non-reasoning)87.9
4Qwen 3 14B (Non-reasoning)87.5
5Gemma 3 27B (IT)87.2
6Phi-485.4
7Qwen 3 8B (Non-reasoning)82.6
8Llama 3.1 8B Instruct79
9Mistral 7B Instruct (v0.3)78.6
10Qwen 3 4B (Non-reasoning)75.4
11gemma-7B75.1
12Qwen 3 0.6B (Non-reasoning)24.6

Interactive version: theaggregate.ai/benchmark?slug=bagel-wikipedia-geographic-distribution · How It Works · Data refreshed daily, snapshot 2026-10-07.