Partly True
A circulating claim that a company is ending development of its proprietary AI model while shifting resources to frontier AI development has been assessed as partly accurate, following a review of 12 primary sources. Elements of the claim are supported by available documentation, while others remain unconfirmed.
The claim asserts two things: that the company's proprietary AI model is being discontinued, and that its efforts have been redirected toward developing a frontier AI model. The two assertions are presented together as a single narrative of a strategic pivot.
Among the 12 primary sources reviewed, some material backs the notion of a strategic shift in AI development direction. However, the evidence does not fully establish both parts of the claim. In particular, whether the proprietary model has been formally discontinued — as opposed to deprioritized, restructured, or absorbed into a broader program — is not clearly documented.
The "frontier AI" framing also carries uncertainty. The sources indicate activity consistent with large-scale model ambitions, but they do not conclusively confirm that the effort is officially designated or branded as a frontier AI program.
No documentation reviewed specifies a timeline for discontinuation of the proprietary model, nor does it confirm whether existing services built on that model will be migrated or retired. The distinction between an internal development pivot and a public cancellation remains unresolved, and the claim's stronger phrasing — "중단 기로" (set to be discontinued) — goes beyond what the sources establish.
The claim is Partly True.