Binance’s newly introduced Binance Intelligence suite puts Binance AI Pro in the spotlight for more advanced users, and a hands-on test by MarsBit author Biteye focused on a simple question: what has actually changed from earlier versions, and can the product do more than answer questions or pull market data.
Binance groups the new lineup into three layers. Binance AI is aimed at general users and is free to use. It can generate briefings on crypto, U.S. equities, and macro markets, while adjusting content based on a user’s habits, preferences, and holdings. AI Pro is built for more advanced users and is designed to turn natural-language trading ideas into automated workflows. Agent OS targets developers and allows external AI agents to connect with Binance data and trading capabilities.
At the launch event, Binance CEO Richard Teng and Vice President of Product Jeff Li said AI Pro lets users describe a trading idea in one sentence, after which the system organizes it into a workflow that can be edited and simulated before going live. Binance summarized that process as Prompt, Test, Trade.
How Binance positions AI Pro inside the new product stack
The launch presentation also showed use cases beyond direct trading. Binance demonstrated automated market daily reports and continuous BTC monitoring with alerts at key price levels. In Biteye’s account, those demos made the product’s role clearer: AI Pro is meant to handle monitoring, organization, and pre-trade setup, not just conversation.
First test: using AI Pro to screen Binance Earn products
Rather than repeat basic checks such as price lookup or data retrieval, the author started with a practical yield scenario. The prompt asked AI Pro how to allocate 50 BTC in Binance Earn to generate returns without taking principal-loss risk, while prioritizing options with relatively high yield and relatively low risk.
According to the test, AI Pro directly pulled current Binance Earn product data and also read the user’s real account balance. Because the author’s account actually held only a small amount of BTC, the system proactively noted that there were not 50 BTC in the spot account and that BTC would need to be transferred in before the allocation could be executed.
On product selection, AI Pro also reflected the user’s stated constraints. Since the prompt explicitly ruled out principal-loss risk, it excluded products such as dual investment that could change the amount of BTC held, and instead prioritized flexible and fixed-term BTC products.
At the time of the test, it found a BTC flexible product with an APR of about 0.26%, along with a special five-day fixed-term offer with APR up to 20%. Biteye wrote that the result was broadly consistent with what later appeared on the Binance Earn page.
Second test: turning a multi-factor trading idea into a strategy
The more important part of the trial was strategy creation. To see whether AI Pro could handle actual trading logic rather than a simple breakout rule, the author used a prompt that combined several timeframes and market conditions in one request.
The prompt asked AI Pro to create a BTCUSDT perpetual long strategy with these conditions: 1-hour and 4-hour price above MA20; 15-minute price reclaiming MA20 with turnover above 1.3 times the average of the previous 20 candles; open interest up more than 5% over the past hour; taker buy/sell ratio above 1; and funding rate below 0.01% as the trigger to open a position. The requested setup used a 10% position, 2x leverage, a 1.5% stop-loss, and a 3% take-profit, with the instruction to generate the strategy first and not send it live.
Biteye said the result was better than expected. AI Pro broke the logic into separate layers: 1-hour and 4-hour conditions for the broader trend, 15-minute conditions for entry timing, turnover for volume confirmation, open interest and taker buy/sell ratio for capital participation, and funding rate as a filter against overheated conditions.
The test also highlighted how the system handled execution details. For example, when the prompt specified a 10% position, AI Pro explained how it interpreted that figure, including whether it referred to notional exposure or margin allocation. It also surfaced parameters that were not explicitly defined, such as isolated margin, one-way position mode, and checking frequency, and asked the user to confirm them before proceeding.
That step matters because it shows the product is not simply converting a sentence into a rigid rule set. It is also identifying settings that could materially change how the strategy behaves once deployed.
Paper trading before any live deployment
After the strategy was confirmed, AI Pro allowed it to be sent to a simulated environment first. The author placed the BTCUSDT perpetual long strategy into paper trading so the system could keep checking live market conditions and see whether the rules would trigger as intended.
The purpose was twofold: to verify that the strategy could trigger normally and to catch any issues in the logic or parameter setup before considering live trading. According to the article, the strategy is currently running in simulation, and a move to live trading would only be considered if performance remains stable and the trigger logic matches expectations.
The author’s takeaway
After the two tests, Biteye said the updated AI Pro left a better impression than expected. The key point was not whether the system could chat, but whether it could watch markets around the clock, retrieve data, wait for conditions, and prepare a strategy for execution.
The article argues that AI is well suited to repetitive trading tasks because people get tired and lose focus, while software can keep monitoring, checking data, and waiting for triggers. In that context, the author linked AI Pro to the broader move by trading tools toward AI-agent-style products.
Biteye also pointed to Binance’s built-in advantage: exchange depth, a large user base, and direct access to first-hand trading data. Many third-party AI tools still need to connect data feeds, accounts, and execution interfaces on their own. Binance, by contrast, can place AI directly inside an existing trading system.
The author’s final conclusion was straightforward: AI is starting to enter the trading process in a more concrete way. Based on this test, Binance has connected natural-language input, strategy generation, parameter confirmation, and paper-trading deployment into a more complete workflow.