Autonomous arbitrage intelligence by people for people. A project by @sushiterminal
Joined September 2026
- Tweets35
- Following5
- Followers666
- Likes16
AI6’s third fee distribution is complete.
The latest token fees have been sent to holders. You can view the payout transactions below:
solscan.io/tx/D7qLvZLXJySUVV…
solscan.io/tx/3sB4fykYpGc8Rm…
solscan.io/tx/37xUJDHH4zVDxr…
solscan.io/tx/3VEusu2oWUBGmj…
solscan.io/tx/2JgPDD3Gy8sCRH…
solscan.io/tx/4sfG7i1Ra8Yxrt…
AI6’s first distribution of arbitrage profits is complete.
This payout came directly from profits earned by the bot’s trades and was distributed to token holders.
solscan.io/tx/4adQvJgwsh85gW…
solscan.io/tx/4yUrvoA61D1zrn…
solscan.io/tx/5orKB9cN2vjUMj…
solscan.io/tx/39CVyvCSYbAzom…
solscan.io/tx/kWahGsRYtqMrtg…
solscan.io/tx/5Hq2eGhN9oePCf…
solscan.io/tx/2Gg9ueBcEuYtuH…
solscan.io/tx/2S2YRJjV8EWzBr…
All on ai6agent.bot
AI6 currently holds an SPCX position, hedged through futures on @Backpack, @armaniferrante .
The bot is running as intended, with strong community support and feedback helping shape what we build next.
ai6agent.bot/dashboard
AI6 distribution update.
More token fees distributed to holders. Latest transactions:
solscan.io/tx/24mF6EM3CAq2ey…
solscan.io/tx/21cLFmYifg7CUC…
solscan.io/tx/3Gjr8vq46HdiVZ…
solscan.io/tx/2dSAiU3cFfNh2G…
solscan.io/tx/vto5gKUyue1FH2…
.@Backpack team and @armaniferrante 👋
We built AI6 using Backpack’s futures platform to hedge positions and capture price gaps between tokenized stocks on Solana.
Built by @sushiterminal, AI6 is autonomous arbitrage intelligence, by people, for people.
Fees generated on @Pumpfun fund the agent, and its profits are shared with token holders.
We’re participating in the @SolanaFndn Stocklana Hackathon and would love to show the Backpack team what we’re building.
Explore the dashboard: ai6agent.bot
Our 6 campaigns are live on our dashboard and the profit has been sent to solana:DKxXdaMC1so182urvrrnhs6V6fGTrttPS8br6JuEpump holders.
AI6’s trading stats are live on the dashboard.
76 tokens traded
$6.5K in total trading volume
36 transactions
See the bot’s open and closed positions, track net returns, and follow its performance, all in one place.
ai6agent.bot/dashboard
We want our product to be fully transparent, being the first autonomous arbitrage bot participating to Stocklana Hackathon, here are our bot past revenue proves and our holder reward distribution.
Here are our bot revenue distributed to holders :
solscan.io/tx/5HCyL1Jtc8HCCD…
solscan.io/tx/3HDSb5CudJsLWw…
And also a SPCX Profit claimed solscan.io/tx/4yLriTipKrLZPG…
ai6agent.bot
Another $4,000 have been sent!
solscan.io/tx/3T8RPiC35REzzR…
More fees generated from trading the token have been distributed to the holders!
Soon, token holders will also be receiving rewards from our arbitrage bot. More exciting things are coming!
solscan.io/tx/4AHLRJPZ8iapcv…
solscan.io/tx/SjwBCzGmVohqdU…
solscan.io/tx/SjwBCzGmVohqdU…
solscan.io/tx/3tURNW5iBwTmhY…
ai6agent.bot
Our rewarding system is now fully transparent on our claiming wallet :
solscan.io/tx/4AHLRJPZ8iapcv…
solscan.io/tx/SjwBCzGmVohqdU…
solscan.io/tx/SjwBCzGmVohqdU…
solscan.io/tx/3tURNW5iBwTmhY…
Follow our agent live on ai6agent.bot
First fees from our token arrived, 50% of it are directly redirected to $AI6 holders and the other 50% are injected into AI6 protocol, you can track live everything through our live dashboard :
ai6agent.bot/dashboard
Here is the first transaction :
solscan.io/tx/56tjaVqz2BVXgx…
ai6agent.bot
AI6 Agent retweeted
I previously shared @ai6agent’s GitHub profile. However, with GitHub’s security measures, recent projects are not publicly displayed on newly created profiles.
I’ll personally publish and manage @ai6agent’s projects through my personal GitHub account, so all of you can access, explore, and read through the work publicly.
My personal GitHub : github.com/sushiterminal
How we monitor 13 trading pairs?
Our system monitors 13 pairs of tokenized equities and corresponding perpetual markets. For each pair, it collects data from two sources: executable swap routes through Jupiter and the live Backpack order book.
First, the system requests a Jupiter quote to buy the token with $3,000. It then requests a reverse quote to sell the same token quantity back into USDC. This allows us to evaluate both the spot entry and the conditions of a potential exit.
On Backpack, we do not rely only on the best price shown in the order book. The system analyzes the available market depth and calculates VWAP for the entire position size. If there is not enough liquidity, the opportunity is not considered executable.
Data freshness is checked separately. A stale Backpack order book cannot produce a valid trading signal, even when the visible spread looks attractive. When a positive opportunity appears, the system takes additional samples to confirm that it is not a temporary anomaly.
The scanning frequency adapts to market conditions. Pairs with more promising readings are checked more often, while inactive pairs are checked less frequently. A failure in one pair does not stop the entire monitor. The system applies a separate delay to that pair and continues analyzing the remaining markets.
Read more bellow :
Why a two leg trade cannot be atomic?
Buying spot through Jupiter and opening a short on Backpack happen on two independent venues. They cannot be combined into a single transaction that either completes entirely or rolls back entirely.
The system first buys the token through Jupiter and waits for finalized Solana confirmation. It then checks the factual change in the wallet balance. The hedge size is calculated from the amount of tokens actually received, not from the amount predicted by the original quote.
This creates a window of directional exposure between the two legs. During that time, the market can move, liquidity can disappear, or the hedge order can fill only partially.
If the complete hedge cannot be confirmed, the system does not report the position as successfully opened. The trade moves to MANUAL_INTERVENTION, allowing the operator to inspect the factual position and select an appropriate recovery action.
The core principle is simple: an explicit incomplete state is safer than a false success. We cannot remove the execution gap between two independent venues, but we can measure it, limit it, and make it visible.
Why seeing a spread is not enough
A visible price difference between two markets may look like an immediate trading opportunity. In practice, the spread shown on a screen is only a snapshot. It does not tell us whether the full trade can be executed at those prices.
Liquidity matters as much as the quoted price. A small amount may be available at the best bid or ask, while the rest of the order fills at worse levels. That is why we calculate the average execution price for the entire position using Jupiter routes and Backpack order book depth.
We also include the costs of entering and exiting both legs. These include Backpack taker fees, expected slippage, Solana transaction costs, priority fees, funding, and a reserve for partial execution. A spread that looks profitable before expenses can disappear once these costs are included.
Freshness matters too. Jupiter quotes and Backpack order books can change within seconds. If market data is stale, the system rejects the signal. A positive opportunity must also remain valid across repeated checks before it can be considered for execution.
The real question is not how large the visible spread is. The real question is how much value remains after liquidity, fees, slippage, timing, and execution risk are accounted for. Only that remaining edge can become a trading signal.
See you soon on ai6agent.bot
