ModelRouterUpdateBillingRule

The billing type. Must be configurable.

The billing configuration, specified as a JSON object.

1. Token-based tiered pricing (token_tiered)
   Applies to Chat models. Billing is tiered based on the token count, with separate rates for standard mode, thinking mode, and cache hits. The JSON format is as follows:
   json { "tiers": [ { "min_tokens": 0, "max_tokens": 32000, "input_price": 2.5, "output_price": 10, "thinking_input_price": 2.5, "thinking_output_price": 10, "cached_input_price": 2.5 }, { "min_tokens": 32000, "max_tokens": 128000, "input_price": 4, "output_price": 16, "thinking_input_price": 4, "thinking_output_price": 16, "cached_input_price": 4 } ] } Field descriptions:
   Field | Type | Required | Description | Unit
   ---|---|---|---|---
   tiers | array | Yes | A required array of tiered pricing objects. | -
   tiers[].min_tokens | integer | Yes | The lower bound (inclusive) of the token count for the current tier. | token
   tiers[].max_tokens | integer | Yes | The upper bound (exclusive) of the token count for the current tier. A value of 0 indicates no upper limit. | token
   tiers[].input_price | number | Yes | The unit price for input tokens in standard mode. | CNY / million tokens
   tiers[].output_price | number | Yes | The unit price for output tokens in standard mode. | CNY / million tokens
   tiers[].thinking_input_price | number | No | The unit price for input tokens in thinking mode. | CNY / million tokens
   tiers[].thinking_output_price | number | No | The unit price for output tokens in thinking mode. | CNY / million tokens
   tiers[].cached_input_price | number | No | The unit price for input tokens on a cache hit. | CNY / million tokens
   thinking_mode_tiers | array | No | Dedicated tiers for thinking mode. The structure is the same as tiers. If a request includes a reasoning_tokens field and this array is not empty, this array overrides the default tiers. | -
   Constraints:
   
   
   
   
   
   
   
   
   
   
   
   
   
   

2. Per-image billing (per_image)
   Applies to ImageGeneration and ImageEdit models. Billing is based on the number of images generated or processed. The JSON format is as follows:
   json { "price_per_image": 0.2 } Field descriptions:
   Field | Type | Required | Description | Unit
   ---|---|---|---|---
   price_per_image | number | Yes | The unit price per image. | CNY / image
   
   
   
   
   

3. Video matrix billing (video_matrix)
   Applies to VideoGeneration and VideoImageGeneration models. Pricing is based on a combination of video resolution and the presence of audio.
   Note: The server automatically converts the matrix field, used by the frontend UI, to the tiers field required by the API.
   The JSON format is as follows:
   json { "tiers": [ { "resolution": 480, "has_audio": 0, "price_per_second": 0.24 }, { "resolution": 480, "has_audio": 1, "price_per_second": 0.24 }, { "resolution": 720, "has_audio": 0, "price_per_second": 0.24 }, { "resolution": 720, "has_audio": 1, "price_per_second": 0.24 } ], "default_price_per_second": 0.24 } Field descriptions:
   Field | Type | Required | Description | Unit
   ---|---|---|---|---
   tiers | array | Yes | An array of video matrix pricing objects. | -
   tiers[].resolution | integer | Yes | The video resolution. Valid values: 480, 720, and 1080. | Pixel height (p)
   tiers[].has_audio | integer | Yes | Indicates whether audio is included. Valid values: 0 (no audio) and 1 (with audio). | -
   tiers[].price_per_second | number | Yes | The unit price per second for this combination. | CNY / second
   default_price_per_second | number | No | The default unit price per second, applied if the request's parameters do not match any item in the tiers array. | CNY / second
   Constraints:
   
   
   
   
   
   
   
   
   
   
   
   

4. Per-duration billing (per_duration)
   Applies to speech recognition (ASR) models. Billing is based on the duration of the audio. The JSON format is as follows:
   json { "price_per_unit": 0.00022 } Field descriptions:
   Field | Type | Required | Description | Unit
   ---|---|---|---|---
   price_per_unit | number | Yes | The unit price per second of audio. | CNY / second
   
   
   
   
   

5. Per-character billing (per_character)
   Applies to speech synthesis (TTS) models. Billing is based on the number of characters in the synthesized text. The JSON format is as follows:
   json { "price_per_unit": 0.8 } Field descriptions:
   Field | Type | Required | Description | Unit
   ---|---|---|---|---
   price_per_unit | number | Yes | The unit price per 10,000 characters. | CNY / 10,000 characters
   
   
   
   
   

6. Flat-rate token billing (token_flat)
   Applies to Embedding, Rerank, MultimodalEmbedding, and MultimodalRerank models. This method uses a single unit price without tiers. The JSON format is as follows:
   json { "input_price": 0.5, "multimodal_input_price": 0.5 } Field descriptions:
   Field | Type | Required | Description | Unit
   ---|---|---|---|---
   input_price | number | Yes | The unit price for text-only input tokens. | CNY / million tokens
   multimodal_input_price | number | No | The unit price for multimodal input tokens. | CNY / million tokens
   
   
   
   
   
   

7. Omni-modal multi-dimension billing (omni_multimodal)
   Applies to omni-modal models, such as ChatFullmodal (for example, qwen3.5-omni-plus). Prices are set separately for different input and output modalities, such as text, audio, images, and videos. The JSON format is as follows:
   json { "text_input_price": 7, "audio_input_price": 53, "image_input_price": 7, "video_input_price": 7, "text_output_price": 40, "audio_output_price": 213, "multi_text_output_price": 0 } Field descriptions:
   Field | Type | Required | Description | Unit
   ---|---|---|---|---
   text_input_price | number | Yes | The unit price for text input tokens. | CNY / million tokens
   audio_input_price | number | Yes | The unit price for audio input tokens. | CNY / million tokens
   image_input_price | number | No | The unit price for image input tokens. | CNY / million tokens
   video_input_price | number | No | The unit price for video input tokens. | CNY / million tokens
   text_output_price | number | Yes | The unit price for text output tokens. | CNY / million tokens
   audio_output_price | number | No | The unit price for audio output tokens. | CNY / million tokens
   multi_text_output_price | number | No | The unit price for text output tokens when the input is multimodal (contains images, audio, or video). | CNY / million tokens
   
   Automatic billing type mapping
   Model type (model_type) | Auto-detected billing type | Key identifying field in pricingConfig
   ---|---|---
   Chat | token_tiered | The tiers field is present.
   ChatFullmodal | omni_multimodal | The text_input_price or audio_input_price field is present.
   ImageGeneration / ImageEdit | per_image | The price_per_image field is present.
   VideoGeneration / VideoImageGeneration | video_matrix | The tiers field is present, and its elements contain resolution.
   ASR | per_duration | The price_per_unit field is present (ASR scenario).
   TTS | per_character | The price_per_unit field is present (TTS scenario).
   Embedding / Rerank / MultimodalEmbedding / MultimodalRerank | token_flat | The input_price field is present.
   
   
   
   
   
   
   
   
   
   
   
   
   
   
   
   
   
   
   
   
   
   

The effective time, in RFC3339 format.

The expiration time, in RFC3339 format.

The status of the billing rule. Use this field to enable or disable the rule.

The version number of the billing rule.

The request ID.

Indicates whether the request is successful.

The error code.

The error message.

The HTTP status code.