API reference/Models & pricing
Models & Pricing
Use these endpoints to fetch the models you can call now, with their context length, capabilities, and prices, instead of maintaining a model table by hand.
These endpoints are public: no API key is needed, and they only return models you can call right now. Send them to the API host, the same base URL you use for model calls.
On this pageexpand_more
Model listlink
An OpenAI-compatible model list. By default it returns text models only, so clients that treat every entry as a chat model work as is.
/v1/modelscurl 'https://api.wokey.ai/v1/models'
curl 'https://api.wokey.ai/v1/models?output_modalities=all'Query parameters
output_modalities- Optional. Filter by output type: text, image, video, all, or a comma-separated combination such as text,image. Defaults to text. An invalid value returns 400 invalid_output_modalities.
Response fields
id- The model ID to send as model.
owned_by- The model's primary vendor. When a model has several vendors, the full list is in providers in the pricing catalog.
name- Display name.
context_length- Context window in tokens. Text models only.
max_completion_tokens- Maximum output tokens per request. Text models only.
architecture- input_modalities and output_modalities are the input and output types the model supports and are authoritative; modality is a display string built from them.
pricing- Text model prices in USD per token, as decimal strings. prompt is input, completion is output, and input_cache_read, input_cache_write and input_cache_write_1h are cache reads and writes; entries a model does not support are omitted.
supported_parameters- Capabilities the model is confirmed to support, such as max_tokens and tools.
{
"object": "list",
"data": [
{
"id": "deepseek-v4.1-flash",
"owned_by": "deepseek",
"object": "model",
"created": 0,
"name": "DeepSeek-V4.1-Flash",
"context_length": 1000000,
"max_completion_tokens": 393216,
"architecture": {
"modality": "text+image->text",
"input_modalities": ["text", "image"],
"output_modalities": ["text"]
},
"pricing": {
"prompt": "0.000000075",
"completion": "0.0000003",
"input_cache_read": "0.0000000015"
},
"supported_parameters": ["max_tokens", "tools"]
}
]
}The sample keeps one model and leaves out some fields.
- /models and /v1/models return the same content.
- Responses carry an ETag. Send If-None-Match and an unchanged list returns 304, so polling is cheap.
- A model that is temporarily at capacity stays in the list; calls to it return 429 until capacity frees up.
Pricing cataloglink
Prices for every model you can call now, in one format for text, image and video. Unlike the model list, it returns every type by default.
/v1/models/pricingcurl 'https://api.wokey.ai/v1/models/pricing?model=deepseek-v4.1-flash'
curl 'https://api.wokey.ai/v1/models/pricing?output_modalities=image,video'Query parameters
output_modalities- Optional. Same values as the model list; defaults to all.
model- Optional. Return one model, matched without regard to case. An unknown model ID returns an empty list.
Response fields
owned_by / providers- owned_by is the primary vendor; providers lists the currently enabled vendors in default priority order, using the vendor IDs the provider request parameter accepts. The price is the same whichever vendor serves the request.
pricing_skus- The list of price entries, in the same shape for every model.
pricing_skus[].meter- What is billed. Text models use input_tokens, output_tokens, cache_read_tokens, cache_write_tokens, cache_write_5m_tokens and cache_write_1h_tokens; images use output_images; videos use output_video.
pricing_skus[].unit / quantity- The unit and amount priced: per 1,000,000 tokens for text, per image, and per second of video.
pricing_skus[].price_usd- The Wokey price in USD, as a decimal string.
reference_price_usd / savings_pct- Text models only: the official reference price and the discount against it, in percent.
pricing_skus[].constraints- When the price applies, such as an image size tier, or a video mode, resolution and duration range.
time_of_day- Only for models priced by time of day. current_tier is the current period (peak or off_peak), peak_windows_utc lists the peak hours in UTC, and tiers gives both periods' prices. pricing_skus always show the current period.
prompt_tiers- Only for models priced by input length. When a request's input (cache reads and writes included) exceeds above_prompt_tokens, every meter uses that tier's prices.
{
"object": "list",
"data": [
{
"id": "deepseek-v4.1-flash",
"name": "DeepSeek-V4.1-Flash",
"owned_by": "deepseek",
"providers": ["deepseek", "opencode", "ollama", "volcengine"],
"object": "model",
"available": true,
"currency": "USD",
"pricing_skus": [
{
"sku_id": "deepseek-v4.1-flash:input_tokens",
"meter": "input_tokens",
"unit": "token",
"quantity": 1000000,
"price_usd": "0.075000",
"reference_price_usd": "0.150000",
"savings_pct": "50"
},
{
"sku_id": "deepseek-v4.1-flash:output_tokens",
"meter": "output_tokens",
"unit": "token",
"quantity": 1000000,
"price_usd": "0.300000",
"reference_price_usd": "0.600000",
"savings_pct": "50"
}
]
}
]
}The sample keeps one model and leaves out some fields.
- Prices can change with the time of day and with supply, so this endpoint is never cached and always returns current prices.
Image and video modelslink
curl 'https://api.wokey.ai/v1/images/models'
curl 'https://api.wokey.ai/v1/videos/models'/v1/images/modelsReturns the image models you can call now: supported endpoints (generation, edits), parameters and their allowed values (parameter_schema), streaming and edit support, and the per-image price (pricing_skus). Usage: Image generation →
/v1/videos/modelsReturns video models with at least one sellable variant: generation modes, resolutions, aspect ratios, duration range, multimodal reference limits, and each variant's per-second price (pricing_skus). Usage: Video generation →