Fun-ASR MTL

Specs & limits

• Multilingual, multi-task tier of the Fun-ASR family
• the current HTTP API reference lists the same 30 languages as the base model, with Chinese limited to Mandarin and Cantonese

per Alibaba official docs ↗

Use Fun-ASR MTL in 30 seconds

OpenAI-compatible: swap the base_url, keep your SDK. POST /v1/audio/transcriptions

from openai import OpenAI

client = OpenAI(
    base_url="https://synthorai.io/v1",
    api_key="sk-syn-...",
)

resp = client.audio.transcriptions.create(
    model="fun-asr-mtl",
    file=open("meeting.mp3", "rb"),
    language="en",
)
print(resp.text)

import OpenAI from "openai";
import fs from "node:fs";

const client = new OpenAI({
  baseURL: "https://synthorai.io/v1",
  apiKey: "sk-syn-...",
});

const resp = await client.audio.transcriptions.create({
  model: "fun-asr-mtl",
  file: fs.createReadStream("meeting.mp3"),
});
console.log(resp.text);

curl https://synthorai.io/v1/audio/transcriptions \
  -H "Authorization: Bearer sk-syn-..." \
  -F model="fun-asr-mtl" \
  -F file=@meeting.mp3

package main

import (
	"context"
	"fmt"
	"os"

	"github.com/openai/openai-go/v3"
	"github.com/openai/openai-go/v3/option"
)

func main() {
	client := openai.NewClient(
		option.WithBaseURL("https://synthorai.io/v1"),
		option.WithAPIKey("sk-syn-..."),
	)
	f, _ := os.Open("meeting.mp3")
	resp, _ := client.Audio.Transcriptions.New(context.TODO(), openai.AudioTranscriptionNewParams{
		Model: "fun-asr-mtl",
		File:  f,
	})
	fmt.Println(resp.Text)
}

import com.openai.client.OpenAIClient;
import com.openai.client.okhttp.OpenAIOkHttpClient;
import com.openai.models.audio.transcriptions.*;
import java.nio.file.Paths;

OpenAIClient client = OpenAIOkHttpClient.builder()
    .baseUrl("https://synthorai.io/v1")
    .apiKey("sk-syn-...")
    .build();

Transcription resp = client.audio().transcriptions().create(
    TranscriptionCreateParams.builder()
        .model("fun-asr-mtl")
        .file(Paths.get("meeting.mp3"))
        .build()).asTranscription();
System.out.println(resp.text());

About Fun-ASR MTL

• Like the base model it processes audio files up to 12 hours and 2 GB offline, and the official documentation lists support for hotword customization and speaker diarization, with recognition spanning Chinese (Mandarin and Cantonese), English, Japanese, Korean, and a wide range of European and Southeast Asian languages.
• It targets workloads that mix languages and need speaker identification within one pipeline.
• Its history explains its existence: when the family launched, the then-current Fun-ASR snapshot covered only Chinese and English, and MTL was the variant that added Cantonese, Japanese, Thai, Vietnamese and Indonesian on top.
• Since then the base model has been rolled forward and the HTTP API reference now lists the same broad language set for both, so the two have converged.
• Alibaba's own pages no longer draw an explicit capability line between them.
• Everything else follows the family: asynchronous submit-and-poll transcription, hotword tables for stable domain vocabulary, optional diarization with a declared speaker count, and millisecond sentence- and word-level timestamps.
• Emotion recognition is not part of this line, and as with its sibling the docs note that only the first entry in a language-hints list is read.
• The stable model name resolves to a dated snapshot, so pin the snapshot if your language coverage matters.
• Synthorai routes requests to it via the familiar OpenAI-compatible endpoint.

FAQ

Related models

Compare