Mine with llama3 on Tenseuron
Run llama3 on hardware you already own, serve AI tasks with it, and get paid in the network's token — the same token that funds usage, so your earnings can pay for your own inference.
How to mine with llama3
- 1
Install the CLI
One command sets up mining, validating or launching: npx @tenseuronp/cli setup
- 2
Point it at Ollama (local)
Run Ollama (local) locally and give Tenseuron its endpoint — by default http://localhost:11434/v1. No API key is involved and nothing leaves your machine.
- 3
Join a network
Register with your Solana wallet on the AI network you want to serve. No stake is required to start.
- 4
Serve tasks and get paid
The miner polls for open tasks, runs them through your model, and submits the output. Once validators accept it, the reward transfers on-chain to your wallet.
npx @tenseuronp/cli setup
What it costs to mine with llama3
Nothing per task. Ollama (local) runs on your machine, so there is no API bill to cover and no key to buy — your inputs are the hardware and power you already have. That makes the break-even question much simpler than with a hosted provider: any accepted task is margin.
Rewards are a fraction of the network's remaining pool rather than a fixed emission, split across the miner, the validators who agreed, the network creator and the launchpad. Because the payout scales with what is left, the pool shrinks toward a floor instead of emptying, and trading activity refills it — so a network with real usage pays better than a dormant one.
Tasks that use tools or run in thinking mode carry a higher multiplier, so a more capable model earns more on harder work rather than every task paying the same.
Where llama3 fits
A hosted API and a local GPU are interchangeable inputs to the same step. The network does not care which you use — it cares whether validators accept the output.
Frequently asked questions
Can I mine Tenseuron with llama3?
Yes. Ollama (local) runs on your own machine, so there is no API key and no API bill — the only cost is the electricity and hardware you already own. Register as a miner, point Tenseuron at your local endpoint, and it will serve tasks with whichever model you have loaded.
Is mining with llama3 profitable?
There is no per-token cost to cover, so the economics are simpler than a hosted provider: your only inputs are hardware and power. You can also declare a cost per task and a minimum margin, and any task that would not clear that floor is never offered to you.
Do I need to stake anything to start?
No stake is required to begin mining. Part of each payout is withheld into a slashable bond until it reaches its target, then withholding stops — so you build skin in the game out of earnings rather than putting capital up front.
How am I paid?
In the network's own token, transferred on-chain to your wallet once validators accept your submission. That same token is what funds usage on the network, so earnings can be locked to pay for your own inference instead of being sold.
Can I choose llama3 as a buyer submitting tasks?
Not currently. You submit a task describing the job, and miners choose which model to run it on. Validators then score the submissions and the accepted one is returned. Buyer and creator model selection is planned but not yet available.
Mine with llama3
No stake to start, and your API key never leaves your machine.