Binance’s Chinese-language account rolled out a new AI product structure under the Binance Intelligence brand. The lineup is split into Binance AI for everyday users, Binance AI Pro for more advanced users, and Agent OS for developers.
The republished piece, written by Biteye, said the author has gone through several generations of Binance AI products, starting with the original Binance AI, then the later ChatGPT plugin integration, and then newer tools tied to trading, data, and agents. With this latest release, the question was pretty direct: what really changed in AI Pro, and can it do actual trading-related work instead of just spitting out market commentary.
Binance split the product stack into three layers
Per the article, Binance Intelligence is divided into three parts.
- Binance AI: made for general users, free to use, and able to generate crypto, U.S. equities, and macro market briefings that adapt to a user’s behavior, preferences, and holdings.
- Binance AI Pro: targeted at more advanced users, with support for turning a natural-language trading idea into an automated workflow.
- Agent OS: built for developers who want external AI agents linked to Binance data and trading functions.
At the launch event, Binance CEO Richard Teng and Vice President of Product Jeff Li presented AI Pro as a tool where a user can explain a trading idea in one sentence, and the system then turns it into a workflow that can be edited, simulated, and moved into live trading only after confirmation.
The article said “workflow” was one of the words repeated most during the presentation. Users can define when to open a position, which indicators to watch, and how to handle size. Then the system breaks that into a strategy. After that, the strategy can be revised, tested in simulation, and only then considered for live deployment. Binance summed that flow up as Prompt, Test, Trade.
The event also showed other uses. Things like automated daily market reports and nonstop BTC monitoring with alerts at key price levels.
First test: using AI Pro to screen Binance Earn products
Instead of redoing basic checks like price lookups or market data queries, the author began with a portfolio allocation case. Since the latest AI Pro can access Binance Earn, the prompt asked how to place 50 BTC into Binance Earn for yield while avoiding principal-loss risk, with preference for higher-yield and lower-risk choices.
According to the article, AI Pro pulled current Binance Earn product data and also checked the user’s real account balance. Because the author’s account held only a small amount of BTC, the system pointed that out plainly, saying: "Your current spot account does not contain 50 BTC. If you really want to carry this out, you would first need to transfer BTC into the account."
On product selection, the system understood the instruction to avoid principal-loss risk and filtered out products such as dual investment that could alter the amount of BTC held. So it favored 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%. The author said those results were broadly in line with what later showed up on the Earn page.
Second test: building a BTCUSDT perpetual long strategy from a complex prompt
The second round zeroed in on what the author considered AI Pro’s core feature: strategy creation. Rather than type in a simple breakout instruction, the prompt mixed together several data points commonly used in derivatives trading to see whether AI Pro could keep the logic straight as the setup got more complicated.
The request was to build a BTCUSDT perpetual long strategy with these rules: both the 1h and 4h timeframes must be above MA20; on the 15m chart, price must reclaim MA20 and trading volume must be greater than 1.3x the average of the past 20 candles; in the last hour, OI growth must be above 5%, the taker buy/sell ratio must be above 1, and the funding rate must be below 0.01% before entry. Position size was set at 10%, leverage at 2x, stop-loss at 1.5%, and take-profit at 3%, with instructions to generate the strategy first and not send it live.
The article said the output was better than expected. AI Pro split the logic into a structured process: the 1h and 4h charts established the broader direction, the 15m chart handled entry timing, trading volume confirmed expanding activity, OI and the taker buy/sell ratio checked whether positioning was moving in step, and the funding rate worked as a filter for overheated conditions.
The author also said the system was careful with execution details. For instance, when the prompt said “10% position size,” AI Pro explained how it interpreted that setting, including whether it meant notional exposure or margin allocation. It also brought up items the prompt had not explicitly defined, such as isolated margin, one-way position mode, and checking frequency, and asked the user to confirm them.
After the strategy was confirmed, AI Pro could send it into paper trading. The author then put the workflow into simulation so it could keep checking conditions against live market data. That made it possible to see whether the setup would trigger the way it was supposed to and whether any part of the logic or parameters needed tweaking. The article said the strategy is currently running in paper trading, and live deployment will be considered only if performance stays stable and trigger behavior matches expectations.
The author’s takeaway
After the two tests, the author said the updated AI Pro made a better-than-expected impression.
The article argued that AI’s edge in trading comes from watching markets around the clock, checking data, waiting for conditions, and carrying out rules without fatigue or distraction. And that, it said, is one reason more trading tools have recently started moving in the AI agent direction.
The piece added that Binance has a clear advantage here because of its market depth, large user base, and direct access to first-hand trading data. Many third-party AI products still have to wire up data feeds, user accounts, and trading interfaces themselves, while Binance can place AI straight inside an existing trading system.
The author’s final conclusion was simple: AI is starting to enter trading in a practical way.