GPT-4.5 Preview API pricing is based on token usage for input and output. Prices are listed per 1M tokens, with lower rates for cached input. Tool-specific features may add per-call fees.
Text tokens
Per 1M tokens
Input$75.00
Cached input$37.50
Output$150.00
Example costs (GPT-4.5 Preview)
TASK
APPROX COST
Creative writing session
~$2.00–$10.00
Complex problem solving
~$5.00–$20.00
Emotional intelligence task
~$3.00–$15.00
Modalities
What the model can accept and produce
TextInput and output
ImagesInput only
AudioNot supported
VideoNot supported
Features
Platform-level capabilities
StreamingSupported
Function callingSupported
Structured outputsSupported
Fine-tuningNot supported
DistillationNot supported
Predicted outputsNot supported
Tools
Tools supported by this model when using the Responses API.
Web searchNot supported
File searchSupported
Image generationNot supported
Code interpreterSupported
MCPNot supported
Computer useNot supported
Snapshots
GPT-4.5 Preview model snapshots ensure stable behavior by locking a specific version. See all available snapshots and aliases below.