Credit and Billable Items¶
MOI uses Credit as the unified unit for account charges. Compute resources, AI services, storage, traffic, Genesis, bring-your-own-model access, and monthly value-added subscriptions are all converted into Credit deductions.
What is Credit?¶
Credit is the platform’s unified billing unit. List-price conversion: 10 Credit = ¥1; top-ups can receive discounts. The Credit badge in the top bar shows the current balance. See Pricing below for unit prices and discounts.
Eight billable items¶
Item |
What is billed |
Unit |
|---|---|---|
Compute resources |
Compute for workflows, data loading/export, SQL, and more; query and task resources have different specifications and prices |
credit / hour |
AI services |
Platform-provided AI capabilities such as agent conversations, knowledge-base Q&A, intelligent orchestration, and system operators |
credit / million tokens |
Data storage |
Data objects in Catalog, including raw data, processing results, and vectors |
credit / GB-month |
Public-network traffic |
Network traffic from imports and exports; outbound is billed and inbound is free |
credit / GB |
Genesis |
Externally available model services, priced by model |
credit / million tokens |
Bring your own model (BYOM) |
Service fee for connecting your own model API to the platform |
credit / million tokens |
Dedicated support |
Dedicated support channels and related services |
credit / month |
Enterprise identity |
Enterprise identity integration such as SSO / SAML |
credit / month |
Notes:
SQL execution is not a separate billable item. SQL from SQL Editor or a connection string uses query compute resources, billed by node uptime under Compute resources.
The first six items are usage-based. Dedicated support and enterprise identity are fixed monthly charges, billed after activation even if unused.
Boundaries among the three AI/model charges¶
AI services |
Genesis |
Bring your own model (BYOM) |
|
|---|---|---|---|
What is sold |
Built-in platform capabilities |
Models (Token as a Service) |
Access services: forwarding, statistics, and management |
Model visibility |
Hidden |
Exposed and priced in Model Marketplace |
You own the model; the platform does not charge its inference cost |
Pricing |
Unified price, without separate input/output rates |
By model, with separate input/output/cache rates |
Fixed service fee per token |
Typical scenario |
Agent conversations, knowledge-base Q&A, intelligent orchestration, system operators |
Calling a model directly with an API Key |
Connecting your own model to workflows and other features |
Quick guide:
AI used through the UI, such as conversations or knowledge-base Q&A → AI services
A model called from code with an API Key → Genesis
Your own deployed model → BYOM
Platform agents and knowledge bases may call models internally, but you are billed under the unified AI services item. Genesis on a bill generally comes from direct model API calls.
How each item is measured¶
Item |
Measurement |
|---|---|
Compute resources |
Node uptime in hours; prices vary by specification |
AI services |
Tokens at a unified price, with the model hidden |
Data storage |
GB-month |
Public-network traffic |
Outbound GB |
Genesis |
Input, output, and cache tokens, priced by model |
BYOM |
Token service fee, excluding inference fees paid to the model provider |
Dedicated support / enterprise identity |
Fixed monthly charge |
Pricing¶
Conversion rate and top-up discounts¶
List price: 10 Credit = ¥1. Larger top-ups generally receive a larger unit-price discount. Available discount tiers and package amounts appear in the top-up dialog.
Credit purchases are generally non-refundable and may expire, for example after one year, subject to the top-up terms. Top up based on predictable usage.
List-price reference¶
Item |
List-price reference |
Notes |
|---|---|---|
Compute resources |
For example, CPU-S 2C4G is about 2.4 cr/h; M/L/XL tiers cost more |
SQL uses query resources billed by uptime. |
AI services |
About 80 cr / million tokens |
Hidden model and unified price |
Data storage |
About 3 cr / GB-month |
Includes raw data, processing results, vectors, and more |
Public-network traffic |
About 16 cr / GB |
Outbound |
Genesis |
See Model Marketplace |
By model, with separate input/output/cache rates |
BYOM |
About 2 cr / million tokens; may still require product confirmation |
Service fee only |
Dedicated support |
About 199 cr / month |
|
Enterprise identity |
About 99 cr / month |
Pricing principles¶
Even at the maximum discount, prices cover cloud-provider costs.
AI services hide the model and use unified pricing so the platform can manage quality and cost.
Genesis follows common model API pricing conventions.
BYOM is decoupled from platform AI and does not fluctuate with model prices.
AI services and BYOM generally do not distinguish input from output; Genesis does.