List your company

Spoken Dialogue on URO-Bench Basic Track

85.82Understanding AccuracyGLM-4-Voice
4.07625.29846.5267.742Sep 24, 2025

Evaluation results

MethodLinks
GLM-4-VoiceSize=9B2025.0985.82———————————61.6369.9
Kimi-AudioSize=7B2025.0983.89———————————53.8854.44
Pretrain+TtTSize=3B2025.0957.63———————————39.345.68
LLaMA-OmniSize=8B2025.0953.71———————————41.9364.05
VITA-AudioSize=7B2025.0952.08———————————51.4554.97
TtTSize=3B2025.0943.39———————————2430.08
Qwen2.5-3B (AR)Size=3B2025.0934.32———————————13.1523.68
SLAM-OmniSize=0.5B2025.0931.55———————————26.4542.2
MoshiSize=7B2025.0918.23———————————24.2136.65
Mini-OmniSize=0.5B2025.0915.01———————————14.829.71
SpeechGPTSize=7B2025.099.26———————————13.3435.5
Qwen2.5-3B (NAR)Size=3B2025.097.22———————————10.1220.01
ChromaLLM Scale=4B2026.01—69.0574.1238.6171.1451.6922.7460.2660.4762.0764.2457.44——
Freeze-OmniLLM Scale=7B2026.01—70.8978.8726.2957.7446.952.8142.5652.2348.755.848.28——
GLM-4-VoiceLLM Scale=9B2026.01—90.9591.0764.4773.859.2830.9357.8280.7763.0778.7669.09——
LLaMA-OmniLLM Scale=8B2026.01—45.6280.6816.0650.6545.133.8944.4464.3658.472.1948.14——
Mini-OmniLLM Scale=0.5B2026.01—5.0732.2023.2525.0602.8230.9929.831.4218.06——
Mini-Omni2LLM Scale=0.5B2026.01—8.140.060.6628.4926.9206.9734.8130.736.4321.31——
SLAM-OmniLLM Scale=0.5B2026.01—12.2666.211.3236.9534.65021.8548.9841.0352.6131.59——