GPT-4 Turbo API pricing: the short answer
OpenAI’s GPT-4 Turbo model page lists $10 per 1 million input tokens and $30 per 1 million output tokens. The displayed pricing is accompanied by the label “Batch API price,” but the retrieved excerpt does not show whether that label identifies the applicable billing mode or is a page control. So the billing channel for these figures—and the standard API rate—could not be verified from the retrieved material. The same page marks gpt-4-turbo and the snapshot gpt-4-turbo-2024-04-09 as shut down, so don’t use the displayed figures as a confirmed quote for a new workload.
These are API usage prices, not a ChatGPT subscription price.
How to interpret the listed rates
The model-page excerpt presents the amounts as per 1 million text tokens: $10 for input and $30 for output. It also displays the words “Batch API price,” but the available excerpt does not establish whether this is a selected pricing mode or a control. Treat the channel as unverified; do not assume these are standard API rates or confirmed Batch rates. The retrieved material also does not verify a standard, non-Batch rate.
For arithmetic only, if those displayed amounts applied to a job, 10,000 input tokens and 2,000 output tokens would calculate as:
- Input: 10,000 ÷ 1,000,000 × $10 = $0.10
- Output: 2,000 ÷ 1,000,000 × $30 = $0.06
- Total: $0.16
This is a conditional illustration, not a verified price or bill estimate. GPT-4 Turbo is marked shut down in the model-page excerpt, and the applicable billing channel for the displayed amounts remains unclear.
Is GPT-4 Turbo available for production?
OpenAI’s model documentation marks gpt-4-turbo and gpt-4-turbo-2024-04-09 as shut down. That status means the listed model should not be treated as available for a new production deployment. The page also lists a 128,000-token context window and a 4,096-token maximum output; these specifications do not change the shutdown status or establish present availability.
Bottom line
The model page lists $10 per million input tokens and $30 per million output tokens, alongside the label “Batch API price.” The retrieved material does not resolve the billing channel or verify a standard API rate. It also marks GPT-4 Turbo and its dated snapshot as shut down, so these figures are not a confirmed basis for budgeting a new production workload.