BAGEL - Xeno-canto Modulation: leaderboard

BAGEL Xeno-canto questions on the modulation pattern of a vocalization, answered from text alone. 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 419 Xeno-canto questions on modulation pattern 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 2033. 13 models tracked. Updated 7 Oct 2026.

Top models

#ModelSettingScoreOverall rank
1Qwen 3 8BNon-reasoning38.2#1890
2Gemma 3 27B (IT)32.5#1341
3Mistral 7B Instruct (v0.3)31.7#6759
4Claude Opus 4.631#114
5Llama 3.1 8B Instruct30.5#4655
6Qwen 3 32BNon-reasoning30.3#983
7GPT-5.430.1#138
8Qwen 3 14BNon-reasoning30.1#1284
9gemma-7B24.1#7115
10Qwen 3 0.6BNon-reasoning23.9#8357
11SmolLM2-360M23.6#9572
12Qwen 3 4BNon-reasoning22.7#2379
13Phi-420.5#1951

No result here: #2 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.

Interactive version: theaggregate.ai/benchmark?slug=bagel-xeno-canto-modulation · How It Works · Data refreshed daily, snapshot 2026-11-22.