Report says DeepSeek and Moonshot used OpenAI model outputs for training
AI · First reported 29 Oct, 02:11 · Updated 05:18

The Information reported that DeepSeek and Moonshot AI used outputs from OpenAI’s GPT-6 Astra to train recent models. Separately, OpenAI said it traced a distillation campaign that began in July to people linked to Moonshot; the supplied coverage does not include a response from either Chinese company.
- The Information reported that DeepSeek and Moonshot AI used GPT-6 Astra outputs to train their latest models.
- OpenAI said it traced a campaign that began in July to individuals linked to Moonshot AI, and said it disrupted the activity by July 28.
- The coverage describes distillation as learning from another model’s outputs to build or improve a competing system; the Chinese firms’ responses are not included.
Summary written by Lux AI from 2 articles
Key facts
- Models named in the report
- OpenAI’s GPT-6 Astra; DeepSeek and Moonshot AI’s latest models
- Campaign start
- July 1, according to OpenAI
- Peak activity cited
- 16,000 requests on July 24 and 25, involving more than 4,000 users
- OpenAI’s stated action
- It said it disrupted the activity by July 28
Source spectrum
Left 0 · Centre 2 · Right 0 of 2 rated sources
Background
- An earlier Lux summary said Alphabet reported $127.6 billion in third-quarter revenue, while investing in AI infrastructure and reporting a second consecutive quarter of negative free cash flow. [aba86b2b-e506-4a6e-8869-0dbe15f400c2]
- A separate earlier Lux summary said Google reported that AI Mode queries currently earn less advertising revenue than comparable traditional searches. [15fc74e8-b526-4278-80ca-3292a1a8d6f8]
All coverage
Coverage over time
LeftCentreRight
Articles per hour since first reported, by side.
Ownership
- Corporation 100% (2)
Share of the 2 sources by owner type.