Skip to main content
GET
Estimate cost

Authorizations

Authorization
string
header
required

An organization API key created in Settings → Team, sent as Authorization: Bearer <key>.

Path Parameters

model_id
string
required

A model id from GET /v1/models.

Query Parameters

duration_seconds
number | null

Video or audio duration in seconds when pricing is per-second. Omit for per-generation models.

Required range: x > 0
quality
string | null

Quality tier when the model prices by quality.

tier
string | null

Resolution tier (e.g. 720p, 1080p) when pricing varies by tier.

resolution
string | null

Output size of a model that names it resolution (Wan InfiniteTalk); priced as its quality.

mode
string | null

Model mode (e.g. extend_video) when pricing varies by mode.

variant
string | null

Speed/quality variant when the model declares one; omitted, the catalog default is priced.

Maximum string length: 64
pricing_params
string | null

JSON object of other priced parameter values, keyed by parameter name (e.g. {"aspect_ratio": "9:16"}). These keys are refused in pricing_params: duration, duration_seconds, generate_audio, mode, multi_prompt, multi_shot, quality, tier, variant.

Maximum string length: 2000
idempotency_key
string | null

Same key the subsequent generate call will send, when you want the quote and the charge to resolve the same supplier.

Maximum string length: 128
reference_image_count
integer
default:0

Reference image count when the model charges per reference.

Required range: x >= 0
mask_image_count
integer
default:0

1 when an edit sends a mask image, for models that charge for it.

Required range: 0 <= x <= 1
first_frame_count
integer
default:0

Opening-image count when an image-to-video model charges for its input image.

Required range: 0 <= x <= 1
character_count
integer | null

Text length in characters for character-priced speech models.

Required range: x >= 0
input_tokens
integer | null

Estimated input tokens for token-priced models.

Required range: x >= 0
output_tokens
integer | null

Estimated output tokens for token-priced models.

Required range: x >= 0
image_quality
string | null

Image quality setting when the catalog prices by quality.

image_resolution
string | null

Image resolution tier when the catalog prices by resolution.

aspect_size
string | null

Aspect ratio or concrete WxH size for matrix-priced image models.

generation_count
integer
default:1

Number of outputs when pricing is per image or batch.

Required range: x >= 1
generate_audio
boolean | null

Include native video audio in the price; defaults on for Veo.

prompt_length
integer | null

Characters in the prompt as it reaches the model: count each reference name as its positional mention (@Image1, @Video1). Needed for the chosen-length estimate on Omni; without it, or when the length sentence would not fit max_prompt_length, only the full price is returned.

Required range: 0 <= x <= 1000000
video_input_durations
string | null

JSON input-video durations in seconds by role, e.g. {"source_video":[5.25]}. Generation measures actual files before charging.

Maximum string length: 6000

Response

Successful Response

allowed
boolean
required

False when your plan cannot use this model; see reason.

balance
number
required
can_afford
boolean
required
credits
number
required

List price in credits before discounts.

credits_after_discount
number
required
model
string
required
billable_duration_seconds
number | null
discount_percent
number
default:0
estimated_credits_after_discount
number | null

For a model settled to the provider's usage (for example Omni, priced for a 10 s allowance): what is charged when you generate for the requested duration_seconds, given prompt_length. The charge then moves to the provider's final cost, never above credits_after_discount.

may_charge_more
boolean
default:false

Actual usage may cost more than the upfront estimate. Explicit customer spending caps still apply.

reason
string | null