Speech-to-Text Question Answering on AlpacaEval
85.86AccuracySPEECH-OMNI-LITE
Evaluation results
| Method | Links | |
|---|---|---|
| SPEECH-OMNI-LITEModel Parameters=32B, Projector-only training=true2026.03 | 85.86 | |
| SPEECH-OMNI-LITEModel Parameters=8B2026.03 | 76.26 | |
| VocalNetModel Parameters=8B2026.03 | 75.76 | |
| Qwen2.5-Omni-7BModel Parameters=7B2026.03 | 71.72 | |
| SPEECH-OMNI-LITEModel Parameters=4B, Projector-only training=true2026.03 | 68.18 | |
| GLM-4-VoiceModel Parameters=9B2026.03 | 65.15 | |
| LLaMA-Omni2Model Parameters=7B2026.03 | 64.14 | |
| Baichuan-Omni-1.5Model Parameters=7B2026.03 | 64.14 | |
| Freeze-OmniModel Parameters=7B2026.03 | 46.73 |