Grok Agents Outpace ChatGPT Agents in Coinbase Trading Volume
Coinbase logs showed that AI agents powered by Grok traded about 40 times the dollar volume of ChatGPT-powered agents during a session. The figure measured aggregate buy and sell value, not the number of trades, suggesting Grok agents placed larger orders on average. Coinbase did not disclose order types, assets, or active agent counts, and the ratio does not indicate latency or profitability.
Coinbase’s session records showed AI agents built on Grok generated about forty times the dollar turnover of those using ChatGPT. The measure combined executed buy and sell value, so it points to larger average order sizes rather than a higher trade count.
Both models are offered by outside developers—Grok from xAI, ChatGPT from OpenAI—and can drive automated agents on Coinbase that turn model output into market or limit orders. User settings, default templates, availability, and paid access levels may shape activity, and Coinbase withheld details on order types, assets, and active agent totals.
The reported volume gap may encourage traders and developers to test Grok-based agents more actively, while exchanges could face pressure to disclose clearer agent-level data. If automated agents grow, retail and institutional market participants might encounter larger automated orders and shifting liquidity, though the metric alone says nothing about safety or returns. Regulators and platform operators may also weigh whether more transparency is needed for model-driven trading tools.