openai/whisper-v3
Whisper large v3 transcribes or translates speech from an audio file, in 99 languages.
Model Inputs
Audio to transcribe (.mp3/.mp4/.m4a/.wav/.webm/.ogg)
Drag & drop a file here, or click to select
Supported formats: MP3, MP4, M4A, WAV, WEBM, OGG (max 100MB)
Task to perform on the audio file. Either transcribe or translate.
Language of the audio file. If translate is selected as the task, the audio will be translated to English, regardless of the language selected. If `None` is passed, the language will be automatically detected. This will also increase the inference time.
Level of the chunks to return.
Maximum speech segment duration in seconds before splitting.
Whether to merge consecutive chunks. When enabled, chunks are merged if their combined duration does not exceed max_segment_len.
Version of the model to use. All of the models are the Whisper large variant.
Result
Pricing
openai/
whisper-v3
Pricing for Synexa AI models works differently from other providers. Instead of being billed by time, you are billed by input and output, making pricing more predictable.
For example, generating 100 requests costs about $0.40.
Check out our docs for more information about how per-request pricing works on Synexa.
| Provider | Price ($) | Saving (%) |
|---|---|---|
| Synexa | $0.004 | - |
| fal | $0.004 | same price |
Readme
Whisper large v3, served through an optimised inference path - same word error rate as the reference implementation at roughly double the throughput.
task selects transcription or translation; translation always targets English regardless of the source language. Set language to the source language, or leave it to be detected. chunk_level, max_segment_len and merge_chunks control how the transcript is segmented in the response.
Accepts mp3, mp4, mpeg, mpga, m4a, wav and webm.