All insights

Pricing and plans

GPT-4o Transcribe Pricing: API Rates, Budget Math, and 2027 Shutdown

OpenAI lists GPT-4o Transcribe at $2.50 input and $10.00 output per 1 million audio tokens. See what the published rates establish, what they don’t establish about per-minute budgeting, and the model’s shutdown date.

Illustration for Gpt 4o transcribe pricing

GPT-4o Transcribe API pricing

OpenAI’s model page lists gpt-4o-transcribe at $2.50 per 1 million input tokens and $10.00 per 1 million output tokens, under the pricing category “Audio tokens.” The page displays dollar-denominated rates. These are API rates for this specific model—not a ChatGPT subscription price. (OpenAI model page)

Production note: OpenAI marks gpt-4o-transcribe as deprecated. Its speech-to-text guide says it will shut down on February 26, 2027 and recommends gpt-transcribe for file transcription. Check the replacement model’s pricing separately; don’t assume these rates carry over. (OpenAI speech-to-text guide)

Published pricing category Listed rate
Audio-token input $2.50 per 1M tokens
Audio-token output $10.00 per 1M tokens

OpenAI lists /v1/audio/transcriptions as the transcription endpoint. The model page describes audio and text input with text output, but its pricing table labels the rates “Audio tokens.” It does not explain there how mixed audio/text requests are categorized or how to interpret each usage quantity for every request. (OpenAI model page)

How to estimate a budget

For a rough calculation using the model page’s two listed audio-token categories:

Illustrative estimate = (input tokens × $2.50 + output tokens × $10.00) ÷ 1,000,000

For example, if a hypothetical request were billed as 100,000 input tokens and 20,000 output tokens in those categories:

  • Input: 100,000 × $2.50 ÷ 1,000,000 = $0.25
  • Output: 20,000 × $10.00 ÷ 1,000,000 = $0.20
  • Illustrative total: $0.45

This example demonstrates the arithmetic; it is not a conversion from audio minutes, nor does it establish that these two quantities cover every charge for a particular request. The official materials cited here list token rates, but do not establish a duration-to-token calculation or a per-minute estimate for this exact model. For a production forecast, verify the usage quantities returned for your request type and test representative audio; don’t transfer a third-party per-minute estimate for gpt-4o-transcribe-diarize to this model.

Check the model before you build

  • File transcription: OpenAI recommends gpt-transcribe as the starting model. gpt-4o-transcribe is deprecated and has a stated shutdown date of February 26, 2027. (OpenAI speech-to-text guide)
  • Mini variant: gpt-4o-mini-transcribe is a different model. Do not treat its pricing as the price for gpt-4o-transcribe. (OpenAI model page)
  • Speaker labels: gpt-4o-transcribe-diarize is another distinct model. Its rates or third-party per-minute estimates do not confirm this model’s price. (OpenAI transcription guide)
  • ChatGPT: The rates above are API usage rates, not a ChatGPT subscription price.

A third-party catalog also lists the $2.50 and $10.00 token rates, but its regional labels are unclear and it lists a separate, higher-priced EU route. The official model page supports the rates shown above; the cited official material does not establish those third-party regional conditions. (OpenAI model page)

Sources

GPT-4o Transcribe Model — OpenAI API

File transcription — OpenAI API

Transcription — OpenAI API

GPT 4o Transcribe: API Pricing — Orq.ai

Contact us