On October 29, 2026, Reuters, citing a report by The Information, said that Chinese AI developers DeepSeek and Moonshot AI relied heavily on outputs from OpenAI's flagship GPT-6 Astra model to train their latest models. The report cited people with direct knowledge of the matter. As of publication, OpenAI, DeepSeek, and Moonshot AI had not responded to Reuters' requests for comment, and the report's claims had not been confirmed by any of the three companies.
Which two models the report names
According to The Information, Moonshot AI's Kimi K3.1 Pro, released last week, and DeepSeek's upgraded V4.1-Flash 1017, launched on October 17, both benefited from distilling Astra. Distillation means training one model on the outputs of a stronger model; leading companies argue that it lets latecomers reproduce advanced capabilities more cheaply and quickly. That is why distillation has become a flashpoint among leading AI companies, against the backdrop of US-China technology competition.
Why this has become a major dispute
This is not the first time OpenAI has spoken publicly about distillation. In a blog post on September 30, OpenAI said it had linked a core part of a distillation campaign that began in July to individuals associated with Moonshot AI. The wording matters: that post described a link between a campaign and individuals, which is not the same as confirming the claims in this report. Earlier this month, Anthropic said its measures designed to prevent rivals from copying Claude's capabilities had reduced data harvested in so-called distillation attacks by about 90% to 99%. Leading companies speaking out about distillation in succession shows the dispute has moved from the technique itself to rules, enforcement of terms, and competitive order.
What to watch next
On the facts, the next things to watch are whether the three companies respond and whether The Information discloses more verifiable detail. Without responses and further evidence, this report should be read as a single-source account, not an established conclusion. For developers and businesses using these two companies' models, the more practical concerns are the compliance statements and service stability that follow, rather than an immediate judgment about the models themselves. Whether the distillation dispute moves from warnings to enforcement of terms or legal proceedings will determine how large its impact becomes.
