Submitted by scikkk 22 VoiceAssistant-Eval: Benchmarking AI Assistants across Listening, Speaking, and Viewing LLMs for Reasoning 13 2