How to Read AI API Pricing

The ToAPIs pricing page turns model prices into practical budgeting and routing decisions. For text models, the key is the input/output token relationship. For image and video models, the critical factors are whether billing is per request or per second and how resolution or duration changes total cost.

• Per-token pricing fits text and multimodal reasoning, per-request pricing fits fixed generation tasks, and per-second pricing fits video workloads.
• Use the market page to choose capabilities first, then use pricing to finalize default and fallback routing.

Decisions Supported by the Pricing Page

• Choose default and fallback models for different workload classes.
• Estimate how resolution and duration change costs for image and video tasks.
• Turn headline model prices into practical budget and routing rules.

Recommended Usage Sequence

1. Shortlist candidate models in the market page first.
2. Use the pricing page to confirm billing units and cost shape.
3. Finish in model guide pages to validate parameters, error handling, and integration details.

GPT-Image-2 Integration Path

For GPT-Image-2, use the model guide and official docs for invocation details, reference image rules, and async task behavior rather than relying on pricing content.

• Open `/model-guide/gpt-image-2` for `/v1/images/generations` usage.
• Use the guide to confirm reference_images, image upload, and task status polling.