VoiceAssistant-Eval

Multimodal voice-assistant benchmark covering listening, speaking, viewing, role-play, audio context, and consistency dimensions.

6rows
unified_scoreprimary metric
2026-05-27sampled

Metadata

ID
voiceassistant_eval
Category
Speech
Release
Unknown

Metrics

Unified Score, Average Score on Listening, Average Score on Speaking, Average Score on Viewing

Latest Results

Rows parsed from the official VoiceAssistant-Eval leaderboard table across listening, speaking, and viewing task groups.

Rank Subject Unified Score Model Match Provenance Sampled
1 Qwen2.5-Omni-7B 36.37 — Imported 2026-05-27
2 Qwen2.5-Omni-3B 30.57 — Imported 2026-05-27
3 Baichuan-Omni-1d5 29.66 — Imported 2026-05-27
4 MiniCPM-o-2_6 28.29 — Imported 2026-05-27
5 mini-omni2 6.8 — Imported 2026-05-27
6 moshika-vis-pytorch-bf16 3.64 — Imported 2026-05-27