Skip to content

Benchmark

MMAU Speech

Scores are percentages.

Models measured
6
Scores recorded
11
Sources
2
IQ weight
0.66

Score distribution

69.8%
Mean
71.5%
Median
4.8%
Std dev
63.8%
10th percentile
75.1%
90th percentile
59.8%
Min
59.8%76.6%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 0.66 for chat models, measured over 5 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
1Gemini 2.5 FlashTest-mini· MMAU (audio) data76.6%
2Gemini 2.0 FlashTest-mini· MMAU (audio) data75.1%
3Gemini 2.0 FlashTest· MMAU (audio) data72.9%
4Gemini 2.5 ProTest· MMAU (audio) data72.7%
5Gemini 2.5 Flash-LiteTest-mini· MMAU (audio) data72.1%
6Gemini 2.5 ProTest-mini· MMAU (audio) data71.5%
7Gemini 2.5 FlashTest· MMAU (audio) data68.3%
8Gemini 2.5 Flash-LiteTest· MMAU (audio) data67.5%
9Phi 4 Multimodal InstructTest-mini· MMAU (audio) data67.3%
10Phi 4 Multimodal InstructTest· MMAU (audio) data63.8%
11Qwen2.5-Omni 7B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim59.8%

Data credits: Data by LLM Stats