Airdrop teams pick recipients in two passes. First they score activity, often with points. Then they remove wallets that look like one person in many. Public write-ups from Arbitrum, Hop, Monad, Grass and Hyperliquid show what each pass looks at, and what stays hidden. This guide covers both, with drops tracked on this site.
Three points to qualify. Activity before the Nitro upgrade counted double.
One point off for a 48-hour burst, and for a low balance with under two contract interactions.
Funder and sweep transfers, community detection, clusters of 20 or more addresses.
Addresses caught in the Hop program and flagged clusters leave the list.
How points programs turn activity into a score
A points program is a scoreboard that the team defines and can redefine. Hyperliquid's documentation says it handed out 1,000,000 points a week for six months, from November 1, 2023 to May 1, 2024. A second phase gave out 700,000 points a week from May 29, 2024 until November 2024. The weekly total was fixed, so your share depended on what every other user did that week.
The same page says the points criteria were updated on a recurring basis. Referrals were built in: affiliates earned 1 point for every 4 points their referred users earned. The page also says Hyperliquid reserves the right to modify previous point distributions at its sole discretion. The page does not state how points convert into tokens. Points measure activity, and they are not a claim on a payout.
Arbitrum used a one-off score instead. Eligibility points were capped at 15, and a wallet needed at least 3 to qualify. Activity before the Nitro upgrade on August 31, 2022 counted double. The inputs included bridging, the number of months a wallet transacted in, transaction volume and smart contract use.
What sybil detection looks at
A sybil is one person controlling many wallets to collect many shares. Detection starts from three kinds of evidence: where a wallet's first funds came from, how wallets move money among themselves, and whether their behaviour repeats.
Arbitrum's public repository shows the method in some detail. It builds a transaction graph, with an edge between sender and receiver for each transaction that moves ether. It also builds a flow graph from funder transactions, the first ether a wallet receives, and sweep transactions, the final transfers out. Large clusters are split with the Louvain community detection algorithm. Funding and sweeping leave a public trail on-chain, so anyone can build a similar graph.
The repository says it flags cluster transfers involving 20 or more addresses, common funding sources and similar activity patterns. Its README shows four example clusters of 56 to 121 eligible addresses each. Inputs included Nansen eligibility lists and address tags, exchange deposit addresses and the blacklists from Hop's program. It gives no total for flagged addresses.
Arbitrum's eligibility documentation adds two behaviour rules that cost a wallet one point. One applies when all of a wallet's activity happened within 48 hours. The other applies when the balance was under 0.005 ETH and the wallet made fewer than two smart contract interactions. Addresses identified as sybils in the Hop program were disqualified outright.
Hop let the community do part of the work. Its airdrop repository says a report must list at least 10 addresses still on the eligible list, and that reports are reviewed first come, first served. The method behind a report has to be well explained and easy to understand.
Monad's results post lists other screens. Every claimer wallet was checked against sanctions lists with the Chainalysis oracle, the claim portal blocked several jurisdictions, and team members were excluded. Those are compliance screens, not sybil filters. On sybils the post says only that the allocation was a balancing act between inclusion and sybil-resistance, and it publishes no rules for the Onchain Users track.
Outside firms appear in these records as suppliers. Nansen supplied data to Arbitrum's process, and Chainalysis supplied Monad's sanctions oracle. We found no standalone Nansen or Chainalysis report on airdrop sybil criteria that we could fetch today, so this guide cites the projects' own material.
How recent drops weighted real use over volume
| Project | What drove the score | Filters or screens disclosed | Not disclosed |
|---|---|---|---|
| Arbitrum | Points up to 15, with 3 needed; early activity counted double | 48-hour and low-engagement deductions, Hop exclusions, graph clustering | Total number of flagged addresses |
| Hyperliquid | Points from protocol activity, from a fixed weekly pot | None on the points page | Conversion of points to tokens |
| Monad | Five tracks; the community track was judged by people | Sanctions check, geo-blocking, team exclusion | Rules for the Onchain Users track |
| Grass | Network Points for bandwidth used, priced far above Uptime Points | No fraud or sybil step described in the Stage 2 post | Any sybil rules |
| Jupiter | A usage and stake snapshot | None found in the proposal | Whether the snapshot ever pays out |
Grass gives the clearest price on real use. Its Stage 2 post says the most heavily weighted factor is the actual use of a user's bandwidth. That earns $0.0049 USDC per Network Point, against $0.00000007 per Uptime Point for time online, a gap of about 70,000 times by our arithmetic. Grass also said 150,000 users received about 90 percent of network traffic, because the network favoured stable, updated nodes in high-demand locations. See our Grass page for the claim status.
Bonuses sit on top of those base rates. The desktop app multiplies Network Points by 5 and the Android app by 10, while referral and sign-up bonuses feed Uptime Points. Grass said 1 million Network Points were awarded across the whole network each day, which makes it another fixed pot. Time online is priced far lower per point.
Monad split 4.73 billion MON across five tracks. Its Monad Community track covered 5,935 accounts sharing 1.670 billion MON. The post describes a human process for it: analysis of Discord, Telegram and X data, a Recognizer app for community input, manual review and a vouching system. The Onchain Users track, built from DeFi activity and NFT holdings, covered 229,433 accounts sharing 1.705 billion.
That works out to about 281,000 MON per community account against about 7,400 per onchain account, roughly 38 times more for the judged group. The figures are averages from the official totals, so they hide the spread inside each track. The Community track claimed 99.7 percent of its allocation and Onchain Users claimed 41.0 percent, and the post does not say why. Our Monad page has the full table.
Jupiter shows the limit of a snapshot. The Net-Zero Emissions proposal says the usage and stake snapshot would be kept. It also says fewer than 29 percent of current stakers voted for the 2024 round. Our Jupiter page records that the DAO voted in February 2026 to postpone the 2026 round. A snapshot exists, but a payout does not, and only a new DAO vote could change that.
Hyperliquid sits at the other end. Its points ran for months with criteria that changed, and our tracker records that no second airdrop or snapshot had been announced as of October 7.
Why volume alone is a weak bet
This section is our analysis, not advice. Volume is the easiest activity to produce, so it is the signal teams have the least reason to trust. Three published mechanisms work against it.
- Filters target its shape. Arbitrum took a point off when all activity fell within 48 hours, and its repository flags clusters of 20 or more linked addresses.
- Pots are fixed. Hyperliquid issued the same number of points each week, so a new wave of farmers lowers every earlier user's share.
- Rules move. Hyperliquid reserves the right to change past distributions, and Jupiter's postponed round shows that a snapshot is not a payout.
The costs of volume are certain: trading fees, gas and market risk are paid up front. The reward is a share of a pool whose formula the team can still change. Monad's Onchain Users track shows the arithmetic. A pool of 1.705 billion MON divided among 229,433 accounts is about 7,400 MON each.
Depth is a different signal. The published formulas reward measured use (Grass), months of activity (Arbitrum) and human-judged contribution (Monad Community). None of them promises a payout for any behaviour, and this page does not either.
What you can and cannot know about eligibility
You can know what the project has published: the criteria, the snapshot date, the official checker and the claim terms. Grass offers a rewards checker for anyone awarded more than 0 Network Points in its eligibility period. Your own wallet history is public on a block explorer, so you can compare it with whatever rules exist.
You cannot know the parts that were never published. Monad published no rules for its Onchain Users track. Hyperliquid's points page gives no conversion to tokens. Arbitrum's repository gives no total of flagged addresses. A project can also change its mind, as Jupiter's vote shows.
Whether a wallet was excluded, and why, is visible only where the team publishes lists or reasons. Among the cases here, Arbitrum and Hop published methods, and Hop accepted community reports.
When a new points program appears, five questions separate the published from the guessed:
- Is the formula published, or only the existence of points?
- Is there a snapshot date, or an open-ended season?
- Does the project say how sybils are handled, and by whom?
- Is there an official checker on the project's own domain?
- Does the project reserve the right to change past distributions?
Treat unofficial eligibility checkers as unverified. A checker that asks for a signature or an approval is a claim-page prompt in disguise. Read our claim checklist before you connect, and our farming guide for the wallet setup that limits the damage.
