Six weeks.
A point to take stock.
Six weeks of incentivized research: room to explore ideas, collect evidence, and decide how to advance Open Intelligence next.
Give exploration room.
Give it a review point.
The work begins with gradient descent algorithms. Contributors explore update rules, evaluations produce comparable results, and confirmation runs test whether apparent advances hold up.
At review, the questions are concrete: which discoveries survived confirmation, what did unsuccessful experiments teach us, and should the next cycle deepen this target or investigate another component?
After six weeks, we will discuss whether to continue unchanged, refine the task, or pause to assess results. A pause would be an explicit decision, not an automatic stop.
Full incentives from day one.
Six weeks, then take stock.
The configured launch burn floor is 0% from the first day. The full miner emission pool is available for verified progress; this is not a guaranteed payout. Without sufficient progress, unearned emission still burns.
- Day 1 onwardCompetition begins with no launch withholding. Miners propose algorithms, evaluations compare them, and confirmed improvements earn rewards.
- First six weeksBuild a research record: discoveries, confirmation results, unsuccessful approaches, and evidence about what the benchmark measures.
- After six weeksReview the results and discuss a possible pause to reevaluate the task and incentives. There is no automatic pause or scheduled 100% burn period; the default remains competition with a 0% launch burn floor.
Follow the work as it unfolds.
The dashboard will begin showing data on September 7, 2026, when the validator goes live on mainnet. Round results and emission records will appear as they are published.