Author: Biteye
The day before yesterday, @binancezh announced a new AI product system, Binance Intelligence, which includes Binance AI for ordinary users, Binance AI Pro for advanced users, and Agent OS for developers.
The author is a long-time user of Binance AI products. From the earliest Binance AI, to the later ChatGPT plugin integration, and then to the past few months when Binance has continuously added features in the trading, data, and Agent directions, basically every time a new version of an AI product comes out, the author tries it.
So this time, after seeing the new AI Pro, what the author most wants to know is: how exactly is it different from before? Can it help me make money 24/7?
🌟What is AI Pro?
According to the official breakdown, Binance Intelligence is mainly divided into three layers:
Binance AI: for ordinary users, free to use, can produce briefings on Crypto, US stocks, and the macro market, and also adjusts content based on users' usage, preferences, and holdings
AI Pro: for more advanced users, can directly describe trading ideas in natural language, and then turn them into an automated Workflow
Agent OS: for developers, allowing external AI Agents to connect to Binance's data and trading capabilities
At the launch event, Binance CEO Richard Teng @_RichardTeng and Product Vice President Jeff Li @taida88 introduced AI Pro: users only need to describe their trading idea in one sentence, and AI Pro can organize it into a Workflow, first modify and simulate it, and after confirmation move it into live trading.
Workflow was also a high-frequency term throughout the launch event. You can directly tell the AI under what conditions to open a position, which indicators to watch, and how to control position size, and it then breaks these conditions into a strategy. After generation, you can continue to modify the strategy, first test it with paper trading, and finally decide whether to run it live. The official summary of this process is Prompt, Test, Trade.
In addition to trading, the launch event also demonstrated scenarios such as automatically generating daily market reports, continuously monitoring BTC, and alerting at key price levels.
After watching the launch event, the author had a rough concept of AI Pro. Next, let's directly look at how it actually performs when put into real needs.
🌟Hands-on test: what can AI Pro do
What the author actually most wanted to verify was how these features perform in real trading scenarios. Are they really as impressive as at the launch event?
After all, for trading AI, being able to chat is no longer anything new. What matters more is whether it can truly understand my needs, call real-time data and account information, and then further help me generate strategies and execute tasks.
So this time the author picked two scenarios that I might really use in daily life: one is wealth management, and the other is trading strategy. The former tests whether it can reduce the time spent screening products, while the latter directly tests the core capability of the new AI Pro.
1️⃣Where should 50 BTC go? First let AI Pro help me choose
The previous Binance AI and ChatGPT plugin could already check market conditions and view data, so this time the author did not really want to repeatedly test these basic capabilities, but instead started with a more practical scenario.
It just so happened that the new AI Pro had already integrated Binance Earn, so I directly asked:
I have 50 BTC and want to put it in Binance Earn to earn yield, but I do not want to bear the risk of principal loss. How would you allocate it? Please prioritize recommending currently higher-yield, lower-risk options.
AI Pro directly calls the current product data from Binance Earn, and can also read my real account balance.
Because my account actually only has a very small amount of BTC, it even proactively reminded me: "You currently do not have 50 BTC in your spot account. If you really want to execute this, you still need to transfer BTC into the account first."
In terms of product selection, it also understood my core requirement, proactively excluded products such as dual investment that could change the amount of BTC, and prioritized BTC flexible and fixed-term products. It found that at the time BTC flexible APR was about 0.26%, and there was also a 5-day special promotional fixed-term product with up to 20% APR, basically consistent with what the author later saw on the Earn page.
2️⃣Trading strategy hands-on test: can it understand my trading logic?
Next was the feature the author most wanted to try - "Create Strategy."
This time I deliberately did not give it a simple "buy on breakout," but instead stuffed in several types of data that I usually use when looking at contracts, to see whether it could still sort out the logic clearly once the conditions became complex.
The prompt I gave it was:
Help me create a BTCUSDT perpetual long strategy: 1h and 4h are both above MA20; 15m reclaims MA20 and turnover is higher than 1.3 times the average of the past 20 candles; open a position when OI growth over the past 1h exceeds 5%, the active buy-sell ratio is >1, and funding rate is <0.01%. Position size 10%, 2x leverage, stop loss 1.5%, take profit 3%. Generate the strategy first, no live trading.
Its performance this round was actually better than the author expected.
It broke down the entire logic quite completely: 1h and 4h determine the major-cycle direction, 15m finds the entry point, turnover confirms volume expansion, OI and the active buy-sell ratio observe whether funds are moving in sync, and Funding further filters out overheated market conditions.
The execution-layer details were also handled quite well. For example, whether "position size 10%" refers to notional position or margin proportion, it proactively explains its own understanding; parameters that were not explicitly written out, such as isolated margin, one-way position mode, and check frequency, are also brought up in advance for the user to confirm.
After the strategy is confirmed, AI Pro can also first put it into paper trading. The author also directly put this strategy into simulated trading and let it continuously check the conditions according to real market conditions. In this way, on the one hand you can see whether the strategy will actually trigger normally, and on the other hand you can also discover in advance whether there are problems with the logic or parameters.
At present, this strategy is already running in paper trading. Later, if its performance is stable and the trigger logic also meets expectations, then consider switching it to live trading.
🌟Final thoughts
After two rounds of testing, the author's impression of the new AI Pro was actually better than expected.
When previously researching AI Trading, the author increasingly felt that AI's advantages are watching the market 24 hours a day, checking data, waiting for conditions, and executing strategies. People get tired and distracted, but AI does not. This is precisely where it is best suited for trading, which is also why more and more trading tools have recently started moving in the direction of AI Agents.
And Binance has a very obvious advantage in doing this: Binance has the best depth, a huge user base, and a large amount of first-hand trading data. Many third-party AIs still have to connect data, accounts, and trading interfaces themselves, while Binance can directly put AI into an existing trading system.
So the biggest impression this AI Pro gave the author is: AI has truly begun to enter the trading era.
Makale / PANews