Multilingual ASR coverage
According to the model card, the 0.6B model supports language identification and recognition for 30 languages and 22 Chinese dialects.
Open Source Model Profile · Qwen
Qwen3-ASR-0.6B
Qwen3-ASR-0.6B is a 938M-parameter Qwen speech-recognition model from Qwen. The model card describes 52-language and dialect coverage with streaming and offline inference.
- Publisher
- Qwen
- Task
- automatic-speech-recognition
- Model type
- qwen3_asr
- License
- apache-2.0
- Library
- Unknown
- Publication status
- Accepted · not indexed
Model overview
Qwen3-ASR-0.6B is published by Qwen as an automatic-speech-recognition model. The captured configuration identifies Qwen3ASRForConditionalGeneration with a qwen3_asr model type and Safetensors metadata reports 938008576 parameters. According to the model card, it is the smaller member of the Qwen3-ASR family built on Qwen3-Omni.
Recorded capabilities
Streaming and offline inference
According to the model card, both family models support streaming and offline unified inference, with streaming described as vLLM-backend only.
vLLM and toolkit support
The model card describes a full-featured inference framework with vLLM batch inference, async serving, and streaming, plus day-0 vLLM model support.
Forced-alignment companion
The model card introduces Qwen3-ForcedAligner-0.6B for word- or character-level timestamps from text-speech pairs.
Use cases in the source record
- Multilingual transcription and language identification across the card-listed languages and Chinese dialects.
- Batch, served, and streaming transcription workflows using the documented qwen-asr transformers and vLLM backends.
Limitations and unknowns
- No independent Ethen evaluation results were extracted; published WER tables mainly describe Qwen3-ASR-1.7B and should not be read as 0.6B scores.
- No context-window, VRAM, latency, or pricing values were extracted from this record.
- Provider state is historical snapshot data, not independently refreshed current availability.
- Training, foundation-model, and evaluation-setup details come from the publisher model card and have not been independently verified by Ethen.
Source and provenance
Source: Qwen/Qwen3-ASR-0.6B
Captured: Unknown. Processed: 2026-09-07T19:34:36.232366+00:00.
Qwen3-ASR Overview Introduction The Qwen3-ASR family includes Qwen3-ASR-1.7B and Qwen3-ASR-0.6B, which support language identification and ASR for 52 languages and dialects. Both leverage large-scale speech training data and the strong audio understanding capability of their foundation model, Qwen3-Omni. Experiments show that the 1.7B version achieves state-of-the-art performance among open-source ASR models and is competitive with the strongest proprietary commercial APIs. Here are the main features: All-in-one : Qwen3-ASR-1.7B and Qwen3-ASR-0.6B support language identification and speech recognition for 30 languages and 22 Chinese…
F001F002F003F004F005F006F007F009F010F011F012F013F015F019F023F026F027F034F037