The LLM Responses API allows you to generate structured responses from the popular LLM platforms. It features advanced parameters you can use to customize requests and emulate specific LLM behavior. To date, ChatGPT, Claude, Gemini, and Perplexity are supported. The API’s endpoints support Standard and Live data retrieval methods, depending on the selected AI platform. Your account is billed for setting a task, and the final price also includes the amount charged by the corresponding LLM.
Pricing by mode
| Endpoint | Mode | Turnaround time * | Billing unit | Price |
|---|---|---|---|---|
LLM Responses
/ai_optimization/*/llm_responses/*
|
Standard queuePOST + GET | Up to 72 hours | Per task | $0.0002 + $0.01 advance payment |
| Live modeSingle POST request | Up to 120 seconds on average | Per task | $0.0006 + price charged by LLM |
Turnaround times are averages and can vary slightly with load and task complexity
FAQ
Questions about LLM Responses API
Why can Standard queue processing take up to 72 hours?
Standard queue processing uses separate task-setting and result-retrieval requests and supports batch submission. The published processing window is up to 72 hours; the queue-based request workflow itself does not explain the length of that window.
What is the difference between the LLM Responses API and the LLM Scraper API?
The LLM Scraper API allows you to scrape the responses from LLM as they are displayed to a user. It is a better fit for monitoring and analyzing what users see in AI search results, while the LLM Responses API is more useful for studying how models generate answers under different conditions.
Why is $0.01 charged when I set a task in the Standard mode?
$0.01 is an automatic prepayment required to execute the task, since the amount charged by the LLM is not known in advance. The final price depends on the price charged by the corresponding LLM’s API, and if it is less than $0.01, the difference is refunded to your account balance. Additionally, if the task is not completed within 72 hours, it is marked as failed, and the $0.01 advance is refunded.
How is the price for LLM Responses API endpoints calculated?
The price of each task consists of two parts: the cost for setting the task and the price charged by the LLM API. The latter depends on multiple parameters, which you can review on the pricing page of the corresponding LLM’s API.
The price for setting a task is $0.0006 in the Live mode and $0.0002 in the Standard mode. So, if you set 1,000 Live tasks and the selected model charges $0.002 per response, the price is calculated as follows:
1,000 × 0.0006 + 1,000 × 0.002 = $2.60
You can learn more about the price calculation in this Help Center article.