List your company

Speech-to-Text Question-Answering on LlamaQ, TriviaQA, WebQ, and OBQA Average

57.6AccuracyDIFFUSPEECH
12.04823.87435.747.526Jan 30, 2026

Evaluation results

MethodLinks
DIFFUSPEECHType=Diff., S→S capability=true2026.0157.6
MinMoType=AR, S→S capability=true2026.0154
Qwen2-AudioType=AR, S→S capability=false2026.0146.2
MoshiType=AR, S→S capability=true2026.0146
Phi-4-MultimodalType=AR, S→S capability=false2026.0143.9
Llama-Omni2Type=AR, S→S capability=true2026.0143.5
DiFFAType=Diff., S→S capability=false2026.0143.3
SpiritLMType=AR, S→S capability=true2026.0120.2
SpeechGPTType=AR, S→S capability=true2026.0113.8