Papers1

Qwen3‑ASR is a family of speech models that hear, understand, and write down speech in 52 languages and dialects, plus they can tell you when each word was spoken.