List your company

Automatic Speech Recognition on Common Voice 15

6.1WERQwen3-Omni (Sep 25)
5.48489.637413.7917.9426Apr 27, 2026May 3, 2026May 10, 2026May 16, 2026May 23, 2026May 29, 2026Jun 5, 2026

Evaluation results

MethodLinks
Qwen3-Omni (Sep 25)Wer=06.12026.046.1—————
AFlamingo3 (NeurIPS 25)Model Category=Audio MLLMs2026.047.4—————
Audio Flamingo 3Backbone=Audio Flamingo 3, Method=Baseline2026.067.4—————
Audio Flamingo 3 + CoATBackbone=Audio Flamingo 3, Method=CoAT2026.067.4—————
Qwen2.5-Omni (Mar 25)Model Category=Omnimodal MLLMs2026.047.6—————
Qwen2.5-Omni-7BBackbone=Qwen2.5-Omni-7B, Method=Baseline2026.067.6—————
Qwen2.5-Omni-7B + CoATBackbone=Qwen2.5-Omni-7B, Method=CoAT2026.067.68—————
Omni-o3Model Category=Omnimodal MLLMs2026.047.9—————
Qwen2-Audio (Jul 24)Model Category=Audio MLLMs2026.048.6—————
Qwen2-Audio + CoATBackbone=Qwen2-Audio, Method=CoAT2026.069.25—————
AudioReaonser (Sep 25)Model Category=Audio MLLMs2026.049.5—————
Gemini 2.5 Pro (May 25)Model Category=Closed-Source APIs2026.049.9—————
GPT-4o (May 24)Model Category=Closed-Source APIs2026.0410—————
Ola (Feb 25)Model Category=Omnimodal MLLMs2026.0413.1—————
DeSTA 2.5 (Jul 25)Model Category=Audio MLLMs2026.0420.1—————
Qwen2-AudioBackbone=Qwen2-Audio, Method=Baseline2026.0621.48—————
FastSLMNumber of Parameters=4.8B2026.01—10.94.55———
Gemini-2.5-Flash2026.01—11.26.09———
Gemini-3.1 Pro2026.04———8.5913.46.78
Phi4-MultimodalNumber of Parameters=5.8B2026.01—7.61————
Qwen2-AudioNumber of Parameters=8B2026.01—8.68————
Qwen3.5-Omni-Flash2026.04———4.253.452.68
Qwen3.5-Omni-Plus2026.04———3.461.952.27
Voxtral-miniNumber of Parameters=4.7B2026.01—10.2————
WhisperNumber of Parameters=1.5B2026.01—9.35.74———