Tokens Postpaid
Overview
SenseNova service supports various flexible billing methods. The billing methods are mutually exclusive. This chapter mainly covers the postpaid billing method for Tokens. For other billing methods, please refer to: TPM Quota Package Prepaid, Tokens Quota Package Prepaid
【Tokens Postpaid】：
Definition: Billing based on the actual usage of model services and the corresponding unit price. Different model services have different unit prices.
Billing cycle: Hourly billing, with charges and billing occurring every hour.
Payment method: Postpaid.
Billing units:
Thousand tokens: 1 thousand tokens = 1000 tokens. Charges are based on the actual number of input and output tokens when calling model services.
Input characters: Each Chinese character, single English letter, and punctuation mark counts as one character.
Call rate limit: The default call specification is used for model services under the postpaid mode. The call rate limits for different model services may vary. Model-related rate limits. The number of requests and tokens consumed within one minute by the customer cannot exceed the RPM and TPM in the default call specification. If any item exceeds the limit, the excess requests will fail, with error code 429.
SenseNova LLM
General
Provides various general models with different parameter sizes and context window lengths, suitable for various application scenarios. The latest version V5 model has been comprehensively upgraded, performing excellently in language, knowledge, reasoning, mathematics, and coding, reaching or surpassing GPT-4 Turbo.
| Service | Billing Item | Unit Price | Billing Item | Unit Price |
|---|---|---|---|---|
| SenseChat-5 Model Call | Input tokens | 0.04 RMB/thousand tokens (limited time offer) | Output tokens | 0.1 RMB/thousand tokens |
| SenseChat Model Call | Input tokens | 0.012 RMB/thousand tokens | Output tokens | 0.012 RMB/thousand tokens |
| SenseChat-32K Model Call | Input tokens | 0.036 RMB/thousand tokens | Output tokens | 0.036 RMB/thousand tokens |
| SenseChat-128K Model Call | Input tokens | 0.06 RMB/thousand tokens | Output tokens | 0.06 RMB/thousand tokens |
| SenseChat-Turbo Model Call | Input tokens | 0.002 RMB/thousand tokens (limited time offer) | Output tokens | 0.005 RMB/thousand tokens |
General - Cantonese Version
Specifically designed to adapt to the conversational habits, slang, and local knowledge of the Hong Kong region, significantly enhancing the ability to understand localized Cantonese conversations.
| Service | Billing Item | Unit Price |
|---|---|---|
| SenseChat-5-Cantonese Model Call | Input tokens, Output tokens | 0.027 RMB/thousand tokens |
SenseNova Multimodal Large Model
Integrates visual and language features, achieving world-leading levels in image-text perception. Ranked first in comprehensive scores in the authoritative multimodal large model benchmark test MMBench.
| Service | Billing Item | Unit Price |
|---|---|---|
| SenseChat-Vision Model Call | Input tokens, Output tokens | Free trial for a limited time |
SenseNova TTS Model
Provides functions such as online speech synthesis and rapid voice cloning.
| Service | Billing Item | Unit Price |
|---|---|---|
| nova-tts-1 Model Call | Input characters | Free trial for a limited time |
SenseNova Embedding Model
Converts input text information into vector representations.
| Service | Billing Item | Unit Price |
|---|---|---|
| nova-embedding-stable Model Call | Input tokens | 0.0005 RMB/thousand tokens |
SenseNova Role-play Model
Supports functions such as personalized character creation, knowledge base construction, long conversation memory, and multi-person group chat. Widely used in scenarios like emotional companionship and IP characters.
| Service | Billing Item | Unit Price |
|---|---|---|
| SenseChat-Character-Pro Model Call | Input tokens, Output tokens | 0.015 RMB/thousand tokens |
| SenseChat-Character Model Call | Input tokens, Output tokens | 0.012 RMB/thousand tokens |