> For the complete documentation index, see [llms.txt](https://artisan-ai.gitbook.io/docs/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://artisan-ai.gitbook.io/docs/designing-the-creator-protocol-architecture-and-interfaces/model-execution-and-token-metering.md).
# Model Execution & Token Metering
*Compute is metered, and usage is monetized.*
In ArtisanAI, AI model inference is not a passive backend operation — it is a token-incentivized, programmable economic activity. The protocol treats each model as a first-class on-chain actor, with usage governed by deterministic rules, resource tracking, and monetization logic.
#### Model Execution Workflow
When a user initiates a content generation request (e.g., by submitting a prompt), the following sequence occurs:
* Model Selection
The request is routed to either a public protocol-curated model or a user-deployed custom model.
* Inference Execution
The selected model processes the input and generates output on distributed or edge compute.
* Usage Logging
The model’s CID, user wallet, execution parameters, and timestamp are logged on-chain.
* Metering & Cost Calculation
Based on complexity and compute consumption, the execution is priced in $ART tokens.
* Settlement & Distribution
$ART tokens are programmatically distributed to the model creator, infrastructure operator, and protocol treasury.
#### Dynamic Token Metering
The metering algorithm adjusts execution cost based on:
* Model Type & Size
* Inference Latency
* Priority Tier
* Storage Access
* Concurrent Load
#### Monetization Rights for Model Owners
Model developers can configure monetization terms via smart contracts:
* Fixed or tiered pricing curves
* Access permissions
* Remix licensing splits
* Revenue streaming to DAO treasuries
#### Transparent Execution Trails
All inference events are:
* Recorded on-chain
* Linked to input & output assets
* Auditable & queryable
* Incentive-linked