Veriwire
← Back to the wire

OpenAI's Astra becomes the main target for Chinese AI labs after Anthropic tightens its defenses

RumorModelOct 29, 2026

According to The Information, DeepSeek and Moonshot AI relied heavily on outputs from OpenAI's GPT-6 Astra to train their latest models, including Kimi K3.1 Pro and V4.1-Flash 1017. The reported distillation shifted from Anthropic's Claude after Anthropic tightened defenses, claiming its new methods cut extractable data by 90 to 99 percent. OpenAI reported a coordinated distillation campaign in late September involving people linked to Moonshot; attackers have adapted to each new defense.

Why it matters

Chinese AI labs have shifted their focus from Anthropic's Claude to OpenAI's GPT-6 Astra as the preferred source for training data, according to a report from The Information. DeepSeek and Moonshot AI reportedly leaned on outputs from OpenAI's flagship model to build their newest releases, including Kimi K3.1 Pro and V4.1-Flash 1017. The technique, known as distillation, requires persuading a model to expose its hidden chain of reasoning — something both OpenAI and Anthropic explicitly prohibit in their terms of service. The reported shift follows Anthropic's efforts to harden its systems against extraction; the company says its newer safeguards sharply reduce the data attackers can pull from Claude. OpenAI disclosed a coordinated distillation campaign in late September that it linked to people associated with Moonshot, and says adversaries have repeatedly adapted whenever a new defense is introduced. Moonshot declined to comment, and DeepSeek did not respond to a request for comment. The episode highlights how frontier labs' defenses shape rivals' behavior.

Receipt № 26351 source · awaiting confirmation ◐

Evidence

1source· awaiting independent confirmation

No score is assigned. Sources and their independence are shown in the citation chain below.

Citation chain · 1 source

01high
the-decoder.comOfficialhttps://the-decoder.com/openais-astra-becomes-the-main-target-for-chinese-ai-labs-after-anthropic-tightens-its-defenses/THE DECODER summarizes The Information's reporting that Chinese labs moved distillation efforts to OpenAI's GPT-6 Astra after Anthropic strengthened Claude's defenses.
PRIMARY
OpenAICompanyAnthropicCompanyDeepSeekCompanyClaudeModelMoonshot AICompanyThe InformationCompanyGPT-6 AstraModelTHE DECODERCompanyKimi K3.1 ProModelV4.1-Flash 1017Model

Related

Oct 29Granite Switch 4.1.1 — Update models like you update softwareOct 28Meta Andromeda: Powering the Next Generation of Ads Retrieval Through Model-System Co-DesignOct 28NVIDIA Nemotron 3.5 Transcribe and Granary v2: Intelligent Multilingual Transcription You Can Prompt and Adapt
Canonical: https://the-decoder.com/openais-astra-becomes-the-main-target-for-chinese-ai-labs-after-anthropic-tightens-its-defenses/