tencent cloud

LLM Service TokenHub

DocumentaçãoLLM Service TokenHubToken PlanPersonal Token PlanPersonal Token Plan Credit Deduction Rules

Personal Token Plan Credit Deduction Rules

Download
Modo Foco
Tamanho da Fonte
Última atualização: 2026-08-20 16:23:24
Traduzido por IA
Credits are adopted as the unified consumption unit for usage within the Token Plan Personal Edition package. This document helps you quickly understand the rules for deducting package usage.

Credit Definition

Credits are the unified consumption billing unit for the Token Plan Personal Edition package, used to quantify the consumption of model resources.

Deduction Rules

Credits consumed per request = (Number of input tokens that cache input × Deduction coefficient for cached input + Number of input tokens that miss the cache × Deduction coefficient for non-cached input + Number of output tokens × Credit coefficient for output) / 1,000,000.

Credit Coefficient

Attention:
The DeepSeek-V4 Official Version [Vendor Direct] model has been launched. Meanwhile, in order to better allocate resources and improve service stability, we will adjust the model pricing strategy in sync with the vendor, introducing a Peak/Off-Peak Pricing Mechanism. The off-peak price will be half of the peak price. Peak hours are 9:00–12:00 and 14:00–18:00 Beijing Time (all other hours are off-peak). The new credit deduction pricing will take effect starting from 00:00, August 17, 2026, Beijing Time.
Peak/Off-Peak Period Determination Rule: The billing period is determined based on the period in which the request is initiated. Specifically, requests initiated during peak hours are billed at the peak price for their entire duration. Requests initiated during off-peak hours are billed at the off-peak price for their entire duration, unaffected by the actual completion time of the request.
Model
Model ID
Condition
Deduction Coefficient for Cache-Missed Input
Output Deduction Coefficient
Deduction Coefficient for Cache Input
Auto model
auto
-
82
237
7
GLM-5.2
glm-5.2
glm-5-2
-
154
482
29
Kimi-K2.6
kimi-k2.6
kimi-k-2-6
-
115
474
20
DeepSeek-V4-Pro 0813 Official Release (Vendor Direct)
deepseek-v4-pro-202606
deepseek/deepseek-v4-pro-0813
deepseek/deepseek-v4-pro
OFF-PEAK
94.286
282.857
3.143
PEAK
188.571
565.714
6.286
DeepSeek-V4-Flash 0731 Official Release (Vendor Direct)
deepseek-v4-flash-202605
deepseek/deepseek-v4-flash-0731
deepseek/deepseek-v4-flash
OFF-PEAK
31.429
94.286
1
PEAK
62.857
188.571
2
MiniMax-M3
minimax-m3
Input length (tokens) ≤ 512k
47
188
10
Input length (tokens) > 512k
94
375
19

Deduction Example

Assume the requested model is Kimi-K2.6, and for a specific request, the cached input is 50k tokens, the non-cached input is 10k tokens, and the output is 0.5k tokens. 2.387 credits are consumed. The detailed calculation is as follows:
Credits consumed per request 
= (Number of input tokens that hit the cache × Deduction coefficient for cached input + 
Number of input tokens that miss the cache × Deduction coefficient for non-cached input +
Number of output tokens × Credit coefficient for output) / 1,000,000
= (50000 × 20 + 10000 × 115 + 500 × 474) / 1000000
= 2.387

Anterior: Token Plan OverviewPróximo: Token Plan Quick Start

Ajuda e Suporte

Esta página foi útil?

comentários