Qwen Alibaba Cloud: Qwen3-ASR-1.7B

Name: Qwen3-ASR-1.7B
Author: Qwen Alibaba Cloud

qwen/qwen3-asr-1-7b

active stt 1 provider

Playground

Qwen's 1.7B open-weight multilingual speech recognition model, transcribing pre-recorded audio with word and segment timestamps and optional speaker diarization, and also streaming over an OpenAI-Realtime-shaped WebSocket.

AuthorQwen Alibaba Cloud

Languages18 languages

Variant models3 IDs

Statusactive

Quick start

Qwen3-ASR-1.7B

Create an API key

Create API key

Make a request

Pick an SDK, request mode, and language. The model and provider route are already filled in.

ElevenLabs SDK streaming STT

Streams microphone or file audio through the ElevenLabs-compatible allmodels route.

import { ElevenLabsClient } from "elevenlabs";

const client = new ElevenLabsClient({
  apiKey: process.env.ALLMODELS_API_KEY,
  baseUrl: "https://api.allmodels.io/el"
});

const connection = await client.speechToText.realtime.connect({
  modelId: "qwen/qwen3-asr-1-7b",
  languageCode: "en"
}, {
  queryParams: {
    provider: {
      "only": [
        "sprag"
      ]
    }
  }
});

// Send audio frames to connection, then handle transcript events.

Pass provider parameters

Pass provider preferences through the compatible SDK request options when you want to pin a route to a provider.

{
  queryParams: {
    provider: {
      only: ["sprag"]
    }
  }
}

Variant Models

These are native model IDs accepted by the author/provider for the same canonical model family.

Qwen3-ASR-1.7B Default

Open-weight checkpoint id published by the Qwen team, which hosting providers rename (Sprag serves it as `rhythm`).

Qwen3-ASR-1.7B

Qwen3-ASR-1.7B (transformers build) Variant

Hugging Face transformers-compatible build of the same weights.

Qwen3-ASR-1.7B-hf

Rhythm (Sprag) Alias

Sprag's product id for this checkpoint, and the only value its transcriptions endpoint accepts for `model`.

rhythm

Providers

Different providers can host the same model. Choose one provider when you need a fixed backend, or let routing select among them.

ProviderSlugPer hourLatencyiModesUptime

Sprag AI 14 params

sprag

$0.075

—

Synchronous · Streaming

Sprag AI

14 documented provider parameters for this route.

Pricing

ModeLanguagePer hour

Transcription Any $0.075

Parameters

file body · string · required

Audio to transcribe. Containers: aiff, aac, wav, flac, m4a, webm, mp3, ogg.

model body · string · required

1 enum value

Values

rhythm

frequency_penalty body · number | null · optional

min -2max 2

language body · string | null · optional

ISO-639 tag, not a language name. ENUM ADDED FROM A LIVE PROBE 2026-09-05: Sprag's own spec leaves this a free string and the model card claims 30 languages plus 22 Chinese dialects, but the API rejects th, hi, pl, cs, fil, el, hu, ro, fa, da, fi, sv, mk and sw with HTTP 400, code `unsupported_language`, and a message enumerating exactly these 18. Thirteen of those rejected codes (all but `sw`, which Sprag never documented) are ones the model card claims. Omit the field to have the language detected.

19 enum values

Values

enzhkojaderuitfresptmsnlidtrviyuearur

num_speakers body · integer | null · optional

Hints the speaker count to the diarizer. Ignored without speaker_labels.

presence_penalty body · number | null · optional

min -2max 2

prompt body · string | null · optional

Context prompt. Accepted (200, probe-verified); effect unmeasured.

repetition_penalty body · number | null · optional

min 0max 2

response_format body · string | null · optional

default "json"6 enum values

Values

jsonverbose_jsontextsrtvtt

speaker_labels body · boolean · optional

Adds a 0-indexed integer `speaker` to each attributed word. Requires response_format=verbose_json (400 otherwise) and `word` in any granularity list sent.

default false

temperature body · number | null · optional

min 0max 2

timestamp_granularities[] body · array | null · optional

Optional. A verbose response carries both arrays without it; a list that omits `word` silently drops the words array and still returns 200.

top_k body · integer | null · optional

min 0

top_p body · number | null · optional

min 0max 1

Supported languages

18 languages

English en

Chinese zh

Yue Chinese (Cantonese) yue

Japanese ja

Korean ko

German de

French fr

Spanish es

Portuguese pt

Italian it

Russian ru

Dutch nl

Arabic ar

Turkish tr

Indonesian id

Malay (macrolanguage) ms

Vietnamese vi

Urdu ur

Features

Speech to text

Transcribes pre-recorded audio through an OpenAI /v1/audio/transcriptions shape on the seeded host.

speech_to_text

Realtime streaming transcription

OpenAI-Realtime-shaped WebSocket carrying base64 PCM16 audio inside JSON text frames and returning incremental transcription deltas.

streaming

Word timestamps

response_format=verbose_json returns words[] with float start and end seconds plus segments[], neither of which the realtime path carries.

word_timestamps

Speaker diarization

speaker_labels=true adds a per-request 0-indexed integer speaker to each attributed word, and requires response_format=verbose_json.

diarization

Caption output formats

response_format accepts json, verbose_json, text, srt and vtt.

caption_formats

Context prompt

A `prompt` form field is accepted for recognition context.

context_prompt

Code

ElevenLabs SDK streaming STT

Streams microphone or file audio through the ElevenLabs-compatible allmodels route.

import { ElevenLabsClient } from "elevenlabs";

const client = new ElevenLabsClient({
  apiKey: process.env.ALLMODELS_API_KEY,
  baseUrl: "https://api.allmodels.io/el"
});

const connection = await client.speechToText.realtime.connect({
  modelId: "qwen/qwen3-asr-1-7b",
  languageCode: "en"
}, {
  queryParams: {
    provider: {
      "only": [
        "sprag"
      ]
    }
  }
});

// Send audio frames to connection, then handle transcript events.