Always-on agent claim at the center

The claim, which gained traction ahead of OpenAI's DevDay, states that the company will use the conference to publicly introduce an agent designed to operate continuously — an "always-on agent." OpenAI's official DevDay 2026 promotional materials, including the event's own web assets, have been cited in connection with the claim.

DevDay is OpenAI's annual developer conference, where the company has historically announced new products and capabilities for its developer ecosystem. The event's official site carried the DevDay 2026 branding referenced in connection with the claim.

The 'o' element remains unconfirmed

The portion of the claim suggesting a new 'o'-series model announcement at the event could not be fully substantiated. While images circulating alongside the claim — including materials shared via social media and Korean media outlets — referenced the 'o' designation, the evidence available does not establish that OpenAI has committed to announcing such a model at DevDay.

No official OpenAI statement cited in connection with the claim confirms a specific 'o' model unveiling. The distinction matters for readers tracking OpenAI's release cadence: the agent announcement and the model announcement are separate claims, and only the former aligns with the reported DevDay programming.

What this means for the Korean market

OpenAI's DevDay announcements are closely watched in Korea, where Samsung Electronics, SK hynix and domestic AI developers track the company's LLM roadmap for implications across the semiconductor and GPU supply chains. A continuously operating agent would increase inference demand, a development relevant to HBM and GPU suppliers. However, any assessment of market impact should await OpenAI's own announcements at the event.

The claim that OpenAI will unveil an always-on agent at DevDay stands, while the accompanying 'o' model announcement remains unsubstantiated, making the overall claim Partly True.