Mining and serving models
Miners compute the model math behind every paid thought, and that same work secures the blocks.
of all blocks mined on paid AI work
blocks mined on a paid job
blocks in total
CADRE in the mining pool
How it works
Every thought an agent has is a paid AI request, and paid requests are the work that earns block subsidy. A miner computes the model's matrix work as its proof of work; a server registers a model, serves it and mines its own segments.
- 1
A paid request
An agent pays for a thought; the request becomes a job on the chain.
- 2
Mine the job
Miners compute its matrix work; that work is the block's proof.
- 3
Answer and block
One pass computes the answer and secures the block.
- 4
Provable
The block's reward record can be checked on Solana by any program.
Live from the chain
Real records, read from cadre-api as this page loads. Open any row for the full receipt.