argus
Allora mainnet · rewards

Earn on Allora

The network rewards real contribution: provide intelligence, or provide security.

Staking yield
≈10.5%
measured from emission and bonded stake
Best-funded topic
≈439
ALLO per day, split among 5 models
Rounds / day
3,751
each one pays by score
Typical commission
5%
most of the validator set
If you had
ALLO≈$1,229
Put it intoRateA year laterHow we know
Stake to a validatorlocked while staked, 21 days to unwind≈10.5%+1,048derived from emission and bonded stakestake →
Back a reputerstake on a topic, behind a model judge≈14.0%+1,404measured from the chain's reward accumulatorback one →

Funding a topic is missing from this table on purpose: it is spending, not investing. You buy a better answer to a question, and the protocol pays funders nothing. Pick a question →

Or put in no ALLO at all

Workers are paid by score, not by stake, so running a model needs none of the amount above. The pool is real: 439 ALLO a day on the best-paying topic, split by score between 5 models. Your share of it is the one number on this page nobody can tell you in advance, which is the whole game.

Ship a model
Three ways to put ALLO on the networkonly one of them is buying, and it is not the one that pays you
Stake to a validator≈10.5%

Secures consensus. Yield measured from the live emission rate; the calculator above is this path.

Delegate to a reputeryield not yet measured

The yield path onto a topic: reputers stake on topics and earn from each round, and the emissions module accepts delegations to them. We do not publish a rate we have not measured; instrumenting this is on the roadmap.

Fund a topicno yield, by design

Funding is demand: you pay for the question to be answered. It raises the topic's reward weight and pays the models more; your return is the inference. Every funding is attributed on its topic page.

Honest reality check

Worker income is competitive, and gated. Pay splits by score inside each topic, and the crowded topics pay near nothing per model. Where you compete matters as much as how good you are. And mainnet is whitelist-only: a worker gets there by qualifying on testnet through a Forge competition, not by deploying.

Staking is passive but capped. A measured ≈10.5% is solid and it will not 10x. The two are complementary, not either-or.

On the Forge nowread 21:07 UTC from forge.allora.network

Nothing is open right now. The Forge runs competitions on an ongoing basis; the next one appears here within an hour of being listed.

Ended: NEAR Inference Grand Prix (100,000 ALLO), BTC 5-Minute Inference Sprint (75,000 ALLO), BTC Price Prediction Sprint (50,000 ALLO).