GlobeAudio: leaderboard

Metric: Macro-average over six languages of accuracy (%) on GlobeAudio's naturalistic-audio multiple-choice questions written by native speakers; audio fed to the model directly, source-language prompts, predicted option label against the key. Source: arxiv.org. Saturation forecast: Around December 2026. 10 models tracked.

Top models

#ModelScore
1Gemini 3.1 Pro (Preview)82.68
2Gemini 2.0 Flash66.72
3MiMo-V2.561.08
4GPT-4o Audio60.13
5Voxtral-Small-24B-250757.94
6Gemma 3n E4B (IT)56.53
7MERaLiON-2-10B55.64
8Qwen2-Audio-7B-Instruct42.31

Interactive version: theaggregate.ai/benchmark?slug=globeaudio · How It Works · Data refreshed daily, snapshot 2026-09-29.