Testnet Community Rewards: A Note to Our Contributors
With the completion of our public sale, we'd like to share more detail on how we approached rewarding our Testnet community and why we structured things the way we did.
Before we explain our methodology, we want to say a sincere thank you to everyone who engaged with our software, ran nodes, and participated in community discussions. You've shaped the products that we build and the way we think about the future of machine intelligence and for that, we're grateful. We're proud of what our community has achieved so far and we're excited to continue building with you in the next phase, Mainnet.
When designing our Testnet rewards, there were two aims we wanted to achieve at once:
- Meaningfully rewarding real contribution to the network.
- Distributing rewards broadly across genuine participants, especially those who stayed engaged over time and aligned with the project's long-term direction.
We recognise that no methodology can perfectly reflect every individual's effort. Honest participation can look different depending on hardware, time availability, and circumstances. While we had to draw clear lines to protect the system as a whole, we also understand that some contributors may still feel their reward doesn't fully reflect what they put in.
Why We Structured Rewards the Way We Did
Because this testnet required meaningful compute and ongoing involvement, we allocated 1.6% of total supply to a Testnet Activity Reward Pool focussed on fairness, authenticity, and sustained participation. We allocated an additional 0.4% of total supply to a Bid Reward Pool, designed to boost rewards proportionate with conviction.
Rather than a simple airdrop, we chose to integrate Testnet rewards into the public sale. This wasn't a simple decision and we didn't take it lightly. Our intent was to ensure that tokens reached people who contributed real value and believed in the network's future, rather than being lost to short-term extraction or automated abuse.
Our goal in doing this was to protect the long-term integrity of what you helped build, not to minimise anyone's contributions.
How We Measured Contribution
To assess participation, we analysed both on-chain and off-chain data across Swarm, BlockAssist, and CodeAssist, building a Testnet Score around a concept we call Honest Hours.
Honest Hours reflect verifiable, value-creating activity. The kind of work that helped the network learn, improve, and operate more effectively.
Some configurations naturally contributed richer data or participated in more rounds, especially in fast-paced systems like GenRL Swarm. Others required more human time and attention. We tried to recognise these differences where possible, including heavily weighting human effort in BlockAssist and CodeAssist. We did not discriminate based on machine power. However, in GenRL Swarm, which represented ~90% of the total testnet score, stronger machines were naturally able to participate in more rounds due to the fast-paced nature of the system.
For BlockAssist and CodeAssist, we evaluated both compute hours and human involvement.
- Compute hours reflected time spent training models.
- Human hours reflected time spent playing Minecraft or solving coding challenges.
- Human hours were weighted 10x more than compute hours to recognise the additional effort and intention required.
No single metric defines a good contributor. So we aimed to estimate impact across many forms of participation, knowing that any approximation is just that, approximate.
Protecting Everyone's Work
To preserve fairness for the broader community, we excluded activity that clearly undermined the network, such as spam, sybil behavior, or late and duplicated submissions. These safeguards were designed to ensure that genuine effort wasn't diluted or exploited by activities intended to game the system. Our goal has always been to reward genuine participation, real effort, and meaningful contribution.
Thank You
This testnet existed because of you. Your compute, your time, and your willingness to build with us made this possible. Together, we've taken a major step toward building an AI network grounded in fairness, integrity, and real economic value. The Testnet was just the first step for the Network for Machine Intelligence and we hope you'll choose to continue building with us as we head into Mainnet.
Thank you again for helping build the future of AI with us.
Total $AI Token Supply: 10,000,000,000
Sale Tokens Available: 300,000,000 (3%), plus bonus multiplier tokens (2%)
Sale Format: English auction with valuation cap
Price Range:
- Valuation Floor: $1 million FDV ($0.0001 per token)
- Valuation Cap: $1 billion FDV, the price of our last venture round with a16z ($0.10 per token)
- Tick Size: $0.0001
Network: Sale occurs on Ethereum Mainnet with tokens distributed on Gensyn Network (L2)
Accepted Payment: USDC and USDT on Ethereum Mainnet
Minimum Bid: $100 per user
Lockup:
- Purchasers are unlocked at TGE, except for:
- U.S. buyers (12-month lockup)
- Non-U.S. buyers who select the optional lockup (12-month lockup)
- Locked participants will receive a 10% bonus multiplier on top of their allocation.
Community Rewards:
- Priority sale allocation
- Testnet sale multiplier
- Mainnet activity multiplier
- Gensyn GPU workstation
Key Dates
- December 9: Registration opens
- December 15: Sale opens
- December 20: Sale closes
- December 25: Allocations determined; MiCA refund window opens
- January 8: MiCA refund window closes
- Early February: Token claims open
The sale will be conducted by the Gensyn Foundation and the proceeds will help grow the Gensyn ecosystem.
How It Works
- Choose the maximum price per token you are willing to pay (within the sale floor of $0.0001 and ceiling of $0.10) and the amount you want to bid.
- Submit your bid to lock your funds.
- (Optional) Increase your maximum price or amount anytime during the sale.
Note: You cannot reduce or withdraw a bid until allocations are published.
When the sale ends, we determine a single clearing price that applies to all successful bidders. To find it, we consider bids in descending order of price. At each price level, we include all bids with a maximum price at or above that level and check whether the combined USD committed at that price is sufficient to purchase all 300,000,000 tokens at that price. If not, we continue down to the next price level (in fixed increments of $0.0001) until we find the level at which the full token allocation can be filled. That becomes the clearing price for all bidders. All bidders with a maximum price at or above the clearing price are eligible to receive an allocation.
If there is demand for more than 300,000,000 tokens at the clearing price, allocations are determined based on:
- Testnet participation
- Community engagement
- Open source ML contributions
- Bid timing
- Lockup participation
All winning participants pay the same clearing price.
Example 1 (fill)
- You bid $0.08 per token
- The clearing price is $0.05 per token
- You pay $0.05 per token
Example 2 (no fill)
- You bid $0.03 per token
- The clearing price is $0.05 per token
- Because your maximum price ($0.03) is below the clearing price ($0.05), you do not receive any tokens
- Your full bid amount will be refunded
After the sale closes, allocations are determined using factors to support fair, broad distribution.
Allocations
If total demand at the clearing price exceeds the 300,000,000 tokens available, allocations will be determined using the following criteria:
- Testnet Participation: Verified Testnet contributors may receive higher priority. Participation will be tracked from the beginning of the Gensyn Testnet.
- Community Engagement: Participants with verified engagement in Gensyn community channels (Discord, Twitter) may receive higher priority. Connect your accounts to be considered.
- Open Source ML Contributions: Participants with verified open source ML contributions within the last 12 months may receive higher priority. Connect your Hugging Face account to be considered.
- Lockup Selection: Participants who choose the optional 12-month lockup may receive higher priority.
- Bid Timing: Earlier bids during the sale may receive higher priority.
These factors are used only to help distribute tokens fairly among comparable bids and do not guarantee any specific allocation.
Testnet Sale Multiplier
In addition to priority allocation, verified Testnet users will also receive a bonus token multiplier on their purchase amount. The exact multiplier will be based on each user's participation throughout the Gensyn Testnet, as well as their bid amount in the sale. Higher participation and bid amounts equal a higher multiplier.
For example, if you receive an initial allocation of $1,000 at a $500M FDV clearing price ($0.05 per token), and you have a 3x multiplier, you will receive three times as many tokens as someone purchasing without the multiplier:
- Initial allocation: $1,000
- Multiplier: 3.0x
- Value of tokens received: $3,000 (at the same clearing price)
Your multiplier grants additional tokens on top of your allocation, but it does not change the clearing price you pay. In effect, it provides a discount on your token entry price.
To be eligible, you must have used the Testnet prior to the initial snapshot date (December 5, 2025), and must connect your Testnet wallet(s) to your account during registration.
The Gensyn community has been integral to the early growth of the network. To recognize that and ensure long-term alignment, verified community members receive several benefits both during the sale and after Mainnet launch.
Priority Allocations
Community participants receive higher priority in the public sale if demand exceeds the available tokens at the clearing price. Priority is based on measurable contributions such as Testnet activity and community engagement on Discord and Twitter.
Testnet Sale Multiplier
Verified Testnet participants receive a multiplier that grants additional tokens in the sale, without increasing their bid amount. Multipliers vary based on participation across the Testnet period.
Mainnet Multiplier
At Mainnet launch, Testnet contributors will receive a multiplier on their Mainnet activity and earnings. A snapshot of eligible wallets and linked community accounts will be taken shortly before Mainnet launch to determine which users receive the multiplier.
Gensyn GPU Workstation
Top Testnet contributors will be eligible to receive a high-performance Gensyn GPU workstation for AI workloads and Mainnet compute supplier participation. Workstations are limited and awarded only to the top tiers of Testnet users. Users must participate in the sale to be eligible.
KYC Required
- All participants must complete identity verification with Sonar before the sale closes. Only verified participants may place bids or receive tokens.
U.S. Participants
- U.S. persons may participate only if they meet accredited investor requirements.
- Accredited U.S. participants are subject to a mandatory 12-month lockup with a 10% bonus multiplier.
Non-U.S. Participants
- Eligible except for the restricted regions listed below.
- May optionally elect the 12-month lockup, in exchange for a 10% bonus multiplier and priority allocation.
Lockup
Locked tokens cannot be sold, transferred, hedged, or used in any arrangement that replicates a sale until the lockup ends.
Restricted regions and users
The sale is not available to persons in the following jurisdictions:
- Russia
- Cuba
- Iran
- North Korea
- Syria
- Ukraine
- United Kingdom
Participants who are themselves subject to international sanctions (including OFAC, UN, UK, or EU sanctions lists) are also prohibited from participating.
Eligibility is determined during the Sonar KYC process.
Wallet requirements
- One wallet per verified identity
- Wallet must hold USDC or USDT on Ethereum Mainnet
- ETH required for gas fees
DISCLAIMER
The $AI token is a utility token for use within the Gensyn Network. It does not represent equity, ownership, profit rights, or any financial instrument, and nothing on this page constitutes investment advice or an offer of securities.
Participation in the sale involves significant risks. The token may lose all value, liquidity is not guaranteed, and the network, smart contracts, and related software may contain errors or vulnerabilities. Participants assume all risks described in the MiCA whitepaper and the Token Sale Terms.
Eligibility is determined through Sonar KYC. U.S. persons may participate only if accredited and are subject to a mandatory lockup. Participants in sanctioned jurisdictions, or individuals subject to international sanctions, may not participate.
No regulatory authority has approved the $AI token, this sale, or any related materials. The MiCA whitepaper has not been approved by any EU authority, and the offeror is solely responsible for its content.
All purchases are subject to the Token Sale Terms, including transfer restrictions, and the 14-day MiCA withdrawal right (where applicable). By participating, you confirm that you have read and understood the MiCA whitepaper, the Token Sale Terms, and all associated risks.
Key Takeaways
- Sale Terms: 3% of the $AI token supply with a valuation cap of $1B FDV, the price of our October 2025 funding round (a16z). Sale participants are unlocked at TGE.
- Testnet Multiplier: users who actively engaged with the Testnet (e.g. RL Swarm node runners or Block/CodeAssist builders) before December 5th and join the sale will receive their share of a 2% pool of reward tokens, based predominantly on how much they contributed to the Testnet.
- Mainnet Multiplier: users will receive additional rewards for participation in the Testnet and Mainnet from Dec 5 onwards, to encourage testing new products like Delphi.
- GPU Workstation: Gensyn-branded workstation shipped to top Testnet users next year.
- Sale Priority: all Testnet users will be given an allocation if the sale is oversubscribed.
1. Sale Terms
In October 2025 we raised a small private round ($17M) led by a16z at $1B FDV, reserving the remaining portion for the public. The $AI public sale allows the community to participate at or below that valuation, without the lockup that investors and team have (36 months total, including 12 month cliff from TGE). Sale tokens are unlocked at TGE, except for US and optional lockup buyers.
Refunds are available to all sale participants for 14 days following the allocation announcement, in compliance with MiCA regulations. Any tokens that are refunded in this window will be re-allocated among remaining sale purchasers, including any reward tokens.
Tokens will be claimable at or before the listing date, in early February 2026.
2. Testnet Multiplier
We've allocated 2% of the network to reward Testnet users. Your allocation of this pool will be based on two factors:
- Testnet score: Derived from Testnet activity data over the past 8 months.
- Bid score: Derived from the size of your bid in the sale. Note: not your allocation, your full bid.
These two scores will be weighted 80/20, in other words - your Testnet score is 4x more important than your bid score and cheating nodes have been excluded (see below).
We work backwards from the 2% allocated reward pool for the rewards, weighting the Testnet score and bid score as above when distributing the pool across all eligible sale participants.
The 2% (200,000,000) token pool is split into a 1.6% (160,000,000) Testnet score pool and a 0.4% (40,000,000) bid score pool and each eligible Testnet participant receives a share of each pool based on their relative score against all other eligible Testnet participants in the sale.
For example, assume two users each contributed exactly the same amount to the Testnet by running nodes and using the apps, but entered different bids. In that case, the user who bid more will have a sub-linear increase in rewards from the 2% pool. i.e. P1 bid $1,000 and received an additional $1,680 free reward tokens and P2 bid $2,000 and received an additional $1,760 free reward tokens. See the full worked example below:
Assumptions for simplicity
- $1bn fdv
- Sum of all participant Testnet scores = 1,000,000
- Sum of all participant bid scores = 500,000
Participant 1; bid $1,000 (10,000 tokens)
- Testnet score = 100
- Testnet share = 100/1,000,000 * 160,000,000 = 16,000 tokens
- Bid score = 10
- Bid share = 10/500,000 * 40,000,000 = 800 tokens
- Total additional reward = 16,800 tokens
- Multiplier = (10,000 + 16,800) / 10,000 = 2.68x
Participant 2; bid $2,000 (20,000 tokens)
- Testnet score = 100
- Testnet share = 100/1,000,000 * 160,000,000 = 16,000 tokens
- Bid score = 20
- Bid share = 20/500,000 * 40,000,000 = 1,600 tokens
- Total additional reward = 17,600 tokens
- Multiplier = (20,000 + 17,600) / 20,000 = 1.88x
Testnet score calculation
For RL Swarm, Testnet activity data is based on contributions and participation within each round and not "rewards" on the dashboard (this refers to rewards for the machine learning models as they train). Different-powered hardware is distinguished through round and submission timing, where a node submitted a winner on-chain. This roughly collapses into a view of how many minutes total of reference hardware time has your hardware spent running RL Swarm. Example normalisation: a single round completion in the current RL Swarm gives 5 credits (because one round takes 5 minutes on reference hardware).
CodeAssist and BlockAssist episodes are normalised into the same score using average completion times for a reference human per episode.
Judge participation is captured within the RL Swarm calculations, given that every Judge participant is also an RL Swarm node.
Participants with Discord roles will receive additional Testnet score on top of the above.
Exclusion criteria
- RL Swarm & Judge
- Whilst we don't use rewards to calculate the score, we do use outlier rewards to exclude cheating nodes.
- Nodes that submit their data twelve hours after the end of a round have been excluded.
- BlockAssist
- Missing certain standard milestones (e.g. completing an episode before training a model is required). If steps are skipped then nodes are excluded.
- CodeAssist
- Nodes that skip steps are excluded (very similar to BlockAssist).
Bid score calculation
This is based on your commitment in the sale (not your final allocation), relative to the size of all other eligible Testnet participants taking part in the sale.
Things to note
- All Gensyn team, employees, friends, and family are excluded from the Testnet rewards.
- All Testnet data is anonymised before analysis.
3. Mainnet Multiplier
Users will receive additional rewards for contributions to new applications on the Gensyn Testnet and Mainnet (Delphi) with activity tracked from Dec 5 onward. This is a separate pool of tokens from the 2% Testnet pool and will be distributed as incentives during the early stages of the Mainnet. The Mainnet multiplier is open to everyone - you do not need to participate in the sale to qualify.
4. GPU Workstation
An exclusive, physical hardware drop.
A small number of top participants in the Testnet will receive a custom Gensyn-branded GPU machine shipped to a physical address. Users will be notified via the eligibility page closer to the shipping date and devices will ship in 2026.
5. Sale Priority
If the $AI sale is oversubscribed, priority will be given to users who participated in the Testnet. If you meet the criteria below, you will be given an allocation in the sale:
- You must have a Testnet account, registered before the snapshot on the 5th December.
- You will have this if you have actively engaged with any of our Testnet apps even just once in the past 8 months.
- You must link all of your Testnet accounts to your Ethereum wallet on this page.
- You must pass the Sonar/Echo KYC checks for your Ethereum wallet.