Prezzi on-demand

Official price is the configured token rate before model discount.

Final usage combines input, output, cache and multimodal token line items.

Official price is the configured token rate before model discount.

Final usage combines input, output, cache and multimodal token line items.

Official price is the configured token rate before model discount.

Final usage combines input, output, cache and multimodal token line items.

Official price is the configured token rate before model discount.

Final usage combines input, output, cache and multimodal token line items.

Official price is the configured token rate before model discount.

Final usage combines input, output, cache and multimodal token line items.

Official price is the configured token rate before model discount.

Final usage combines input, output, cache and multimodal token line items.

Official price is the configured token rate before model discount.

Final usage combines input, output, cache and multimodal token line items.

Official price is the configured token rate before model discount.

Final usage combines input, output, cache and multimodal token line items.

Official price is the configured token rate before model discount.

Final usage combines input, output, cache and multimodal token line items.

Official price is the configured token rate before model discount.

Final usage combines input, output, cache and multimodal token line items.

Actual cost = text input tokens × $5/1M + text output tokens × $10/1M + output image tokens × $32/1M.

Cached text and image input tokens are $1.25/1M and $2/1M when reported by the provider.

Listed prices are the single-output pre-run reservation ceilings shown by the model playground; output-image and text-output reservations scale with the image count (n).

output_format "svg" adds the tokens of the text model that writes the SVG (svg_model, default gpt-5.4-mini), billed at that model's own rates listed below. Its reservation covers the author's maximum output (128K tokens) and settles down to the reported usage.

The final charge settles down from that ceiling using the provider's actual token usage. Use View for a request quote.

Actual cost = text input tokens × $5/1M + text output tokens × $10/1M + input image tokens × $8/1M + output image tokens × $32/1M.

Cached text and image input tokens are $1.25/1M and $2/1M when reported by the provider.

Listed prices are the single-output pre-run reservation ceilings shown by the model playground; output-image and text-output reservations scale with the image count (n).

Edit rows include the reservation ceiling for one input image; additional input images increase the reservation.

The final charge settles down from that ceiling using the provider's actual token usage. Use View for a request quote.

Actual cost = text input tokens × $5/1M + text output tokens × $10/1M + output image tokens × $32/1M.

Cached text and image input tokens are $1.25/1M and $2/1M when reported by the provider.

Listed prices are the single-output pre-run reservation ceilings shown by the model playground; output-image and text-output reservations scale with the image count (n).

output_format "svg" adds the tokens of the text model that writes the SVG (svg_model, default gpt-5.4-mini), billed at that model's own rates listed below. Its reservation covers the author's maximum output (128K tokens) and settles down to the reported usage.

The final charge settles down from that ceiling using the provider's actual token usage. Use View for a request quote.

Actual cost = text input tokens × $5/1M + text output tokens × $10/1M + input image tokens × $8/1M + output image tokens × $32/1M.

Cached text and image input tokens are $1.25/1M and $2/1M when reported by the provider.

Listed prices are the single-output pre-run reservation ceilings shown by the model playground; output-image and text-output reservations scale with the image count (n).

Edit rows include the reservation ceiling for one input image; additional input images increase the reservation.

The final charge settles down from that ceiling using the provider's actual token usage. Use View for a request quote.

The displayed rate is an estimate derived from video tokens.

The final charge uses actual completion tokens after the task finishes.

Per-second rates assume no reference-video input and derive from $17.3875 per 1M video tokens.

Requests with reference videos bill every token — output plus reference video — at $10.4 per 1M video tokens; use View for an exact quote.

The final charge uses actual completion tokens after the task finishes.

Actual cost = text input tokens × $5/1M + text output tokens × $10/1M + output image tokens × $32/1M.

Cached text and image input tokens are $1.25/1M and $2/1M when reported by the provider.

Listed prices are the single-output pre-run reservation ceilings shown by the model playground; output-image and text-output reservations scale with the image count (n).

The final charge settles down from that ceiling using the provider's actual token usage. Use View for a request quote.

Actual cost = text input tokens × $5/1M + text output tokens × $10/1M + input image tokens × $8/1M + output image tokens × $32/1M.

Cached text and image input tokens are $1.25/1M and $2/1M when reported by the provider.

Listed prices are the single-output pre-run reservation ceilings shown by the model playground; output-image and text-output reservations scale with the image count (n).

Edit rows include the reservation ceiling for one input image; additional input images increase the reservation.

The final charge settles down from that ceiling using the provider's actual token usage. Use View for a request quote.

Actual cost = text input tokens × $5/1M + text output tokens × $10/1M + output image tokens × $32/1M.

Cached text and image input tokens are $1.25/1M and $2/1M when reported by the provider.

Listed prices are the single-output pre-run reservation ceilings shown by the model playground; output-image and text-output reservations scale with the image count (n).

The final charge settles down from that ceiling using the provider's actual token usage. Use View for a request quote.

Actual cost = text input tokens × $5/1M + text output tokens × $10/1M + input image tokens × $8/1M + output image tokens × $32/1M.

Cached text and image input tokens are $1.25/1M and $2/1M when reported by the provider.

Listed prices are the single-output pre-run reservation ceilings shown by the model playground; output-image and text-output reservations scale with the image count (n).

Edit rows include the reservation ceiling for one input image; additional input images increase the reservation.

The final charge settles down from that ceiling using the provider's actual token usage. Use View for a request quote.