Skip to content

Benchmark

VoiceBench Avg

Scores are percentages.

Models measured
9
Scores recorded
9
Sources
2
IQ weight
0.00

Score distribution

77.8%
Mean
85.5%
Median
13.9%
Std dev
58.1%
10th percentile
91.4%
90th percentile
56.5%
Min
56.5%91.6%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 0.00 for chat models, measured over 7 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
1Qwen3.8 Omni Flash· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim91.6%
2Inkling FP4· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim91.4%
3Inkling Small· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim90.1%
4Qwen3 Omni 30B A3B Thinking· VoiceBench88.8%
5qwen3-omni-30b-a3b-instruct· VoiceBench85.5%
6Qwen2.5-Omni 7B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim74.1%
7Phi 4 Multimodal Instruct· VoiceBench64.3%
8Nemotron 3 VoiceChatV1· VoiceBench58.1%
9GLM-4-Voice· VoiceBench56.5%

Data credits: VoiceBench (NUS HLT Lab), Apache-2.0 (licence) · Data by LLM Stats