gpt-image-2 API, priced per call
OpenAI's image model, strongest at text inside the image and at editing. Priced by the size and quality you ask for.
OpenAI · image model
From
$0.159
per image
gpt-image-2 prices
Prepaid prices, in dollars. A range means the price depends on something only the run decides, such as output size. You are charged what the job actually cost.
- gpt-image-2per image$0.159–$0.24
Every model and unit, machine-readable: GET /api/pricing
Call it from your agent
Connect the hosted MCP server once, then ask in plain words. The agent calls gpt-image-2, waits for the result and hands you the file with its cost.
claude mcp add --transport http aitachyon \
https://aitachyon.com/api/mcp \
--header "Authorization: Bearer ait_your_key"“Use gpt-image-2 to make a poster that reads 'Open late on Fridays'.”
Or from your code
One HTTP call with a bearer key. Video returns a job to poll; images and voice return the file.
curl https://aitachyon.com/api/images/generate \
-H "Authorization: Bearer ait_your_key" \
-H "Content-Type: application/json" \
-d '{"projectId":"<project id>","prompt":"<your prompt>","provider":"openai"}'Why run gpt-image-2 on Aitachyon
- Priced per call in dollars, shown before you run it
- Every job itemised with the model it ran on and what it cost
- A render that fails is refunded to the cent, automatically
- Prepaid balance that never expires; no subscription
- The same key works over HTTP and in Claude Code or Cursor through MCP
gpt-image-2 API questions
On Aitachyon, gpt-image-2 starts at $0.16 per image on a prepaid balance. Every price is listed in dollars before you run the call, and the job shows its real cost afterwards.
No. You top up a prepaid balance and pay per call. The balance does not expire, and you can set an automatic top-up that only runs under a line you choose.
The amount it took is refunded to your balance automatically, to the cent. You never pay for a render you did not get.
Yes. Aitachyon runs a hosted MCP server: one command connects it, then you ask for what you want in plain words and the agent calls the model, waits for the render and hands you the file with its cost.
Other models, same key
All models on one page: /models. Getting started: /developers.