Rules frozen in advance
Challenges, experimental protocols and the scoring framework are published before the submission window opens.
FROZEN RULES01 / REGISTRATION
Enter your team, accept the rules. The first field you fill is the first step of the competition.
Register for the challengeBuilding the molecular scene…
PRINCIPLES & AWARDS
Core rules are published and frozen before submissions open, so different models and design strategies are judged under the same conditions.
Challenges, experimental protocols and the scoring framework are published before the submission window opens.
FROZEN RULESEvery selected design runs through the same standardized assays, so results can be compared directly.
UNIFIED PROTOCOLDesign identity is kept apart from the lab that runs the assay, removing human bias from both the process and the readout.
DOUBLE-BLINDAwards cover overall leaderboard performance, outstanding projects and the winner of each challenge.
Cash awards
May receive an investment offer, subject to due diligence
Provided by Megalaxy Lab; volume announced as the competition progresses
COMPETITION TIMELINE
From October 2026 to April 2027, five challenges are released in two rounds, each followed by one standardized validation campaign.
Runs alongside design submission for round one.
Three challenges are released, submitted and confirmed on the dates below.
Two challenges are released, submitted and confirmed on the dates below.
Results published in April 2027.
CHALLENGE BRIEFS
Five challenges, released in two rounds. Each centers on one real biological target and a specific design problem.
Design a condition-responsive EGFR binder.
Design a condition-responsive TNF-α binder.
Design a GLP1R agonist.
Full briefs, submission requirements and scoring rules are published with each challenge.
COMPETITION REGISTRATION
The 2026 edition is open to teams: research groups, AI teams, biotech companies and university labs anywhere in the world. Any model, any design strategy.
Each challenge comes from a real drug-discovery program. Beyond baseline binding, designs are tested across multiple dimensions, including species cross-reactivity, thermal stability, nonspecific binding, condition-selective binding, cell-based receptor activation, and blocking activity.
EVALUATION SYSTEM
The inaugural edition tests the metrics that matter most in protein design, assessing how well a complete design system generalizes, how often it succeeds and how efficiently it works.
How well the design binds its target.
How well the design withstands temperature.
Binding to the orthologous target from the species used in preclinical testing.
No detectable binding to proteins other than the target.
Binding affinity that changes with pH.
Activation, blockade or other functional activity measured in cells.

WHERE VALIDATION HAPPENS
All experimental validation is planned to take place at Megalaxy Lab. Protein expression, purification, binding assays, cell function and organoid work sit on one traceable pipeline, from sample to data.
The 2026 edition is open to teams: research groups, AI teams, biotech companies and university labs anywhere in the world. Any model, any design strategy.
Alongside the sequences, tell us the molecule class, the expression system you want used, and the AI models behind the design.
No. Round one releases the first three challenges, round two the last two. Full requirements, dates and scoring rules are published when each challenge opens.
Selected designs go to Megalaxy Lab, where they are expressed and assayed on the same automated platform.
Yes. The winning team receives up to ¥1,000,000 in cash, and teams on the leaderboard may be offered investment.