MMAU-Pro - Multi-Audio: leaderboard

Audio-language models answer human-written questions about real-world speech, sound and music clips of up to 10 minutes; the score is accuracy on the 430 questions over several clips, from the authors' runs. Unlike the original MMAU, clips are long, can be mixed or multiple, and include spatial audio, voice chat and open-ended questions.

Metric: Accuracy (%). Source: arxiv.org. Saturation forecast: Estimated already saturated. 18 models tracked. Updated 17 Nov 2026.

Top models

#ModelScoreOverall rank
1Ming-Lite-Omni-1.537.4#2348
2Baichuan-Omni-1.528.8#3064
3Audio Flamingo 326
4Qwen2.5-Omni-7B24.3#1883
5Audio Reasoner22.6#3028
6Mellow20.8
7GAMA20.2#8292
8BAT20.2
9DeSTA2.5-Audio19.8#2076
10Gemma 3n E4B (IT)19.6#2432
11SALMONN-13B17.4
12Audio Flamingo 215.5
13DeSTA2-8B-beta13.2
14Phi-4 Multimodal Instruct11.4#4209
15Gemma 3n E2B (IT)11.4#4283

No result here: #1 Claude Fable 5.2, #2 GPT-6.2 Astra, #3 Gemini 4 Carbon.

Interactive version: theaggregate.ai/benchmark?slug=mmau-pro-multi-audio · How It Works · Data refreshed daily, snapshot 2026-11-17.