Chinese AI labs distilled OpenAI's GPT-6 Astra for new models, report says
Moonshot AI and DeepSeek used outputs from OpenAI's GPT-6 Astra to train their latest artificial intelligence models, according to a report by The Information.
Chinese artificial intelligence firms Moonshot AI and DeepSeek trained their newest models using outputs from OpenAI's GPT-6 Astra, The Information reported. Moonshot distilled Astra during the development of Kimi K3.1 Pro, which was released the week prior, according to three people with direct knowledge of the matter. Two people with direct knowledge also said DeepSeek's V4.1-Flash 1017—an Oct. 17 upgrade to a model first released in September—benefited from distilling Astra.[1][2][3][4]
The reported shift to OpenAI's model follows defensive measures by Anthropic, whose Claude models previously served as the primary target for Chinese distillation campaigns. In February, Anthropic stated it discovered distillation operations by Moonshot, DeepSeek, and MiniMax utilizing roughly 24,000 fraudulent accounts. Anthropic implemented tighter usage defenses in October, which it reported reduced data harvesting by 90% to 99%.[2][3]
OpenAI reports that GPT-6 Astra scored 57.9% on a benchmark test, compared with 55.8% for Claude Fable 5.1 and 37.3% for GPT-5.6 Sol, alongside completing tasks 1.9 times faster than GPT-5.6 Sol on Mind2Web. Separately, Artificial Analysis released Intelligence Index v5 on Oct. 28, increasing its private benchmark share from 45% to 60%, after which Kimi K3.1 Pro's score dropped from 55 to 46. Moonshot did not comment and DeepSeek did not respond to requests for comment.[2][3]
Key facts
- The Information reported that Moonshot AI distilled OpenAI's Astra model for Kimi K3.1 Pro and DeepSeek did so for V4.1-Flash 1017.
- DeepSeek's V4.1-Flash 1017 was released on Oct. 17 as an upgrade to its original September model.
- Anthropic reported finding February distillation campaigns by DeepSeek, Moonshot, and MiniMax that used approximately 24,000 fraudulent accounts.
- Anthropic introduced defenses in October that it said cut data harvested by distillers by 90% to 99%.
- OpenAI reported GPT-6 Astra reached 57.9% on a benchmark, compared to 55.8% for Claude Fable 5.1 and 37.3% for GPT-5.6 Sol.
- Artificial Analysis increased the private share of benchmarks in its Intelligence Index v5 from 45% to 60% on Oct. 28, and Kimi K3.1 Pro's score fell from 55 to 46.
- Moonshot did not comment and DeepSeek did not respond to requests for comment regarding the report.
Sources · 4 sources
- RE
Reuters@ReutersPost on X ·
DeepSeek, Moonshot used OpenAI's GPT-6 Astra to train latest models, The Information reports https://t.co/pICutna1Ah
Open source - CH
Chubby♨️@kimmonismusPost on X ·
As expected, Kimi 3.1 Pro distilled US frontier models to train their models. This time they distilled GPT-6 Astra. The damage to Kimi’s reputation keeps growing. Via the information https://t.co/loSXL7Ruft
Open source - AG
AGTP@AGTPinsightsPost on X ·
Chinese AI labs DeepSeek and Moonshot AI just turned to OpenAI's GPT-6 Astra as a distillation target. Here's what you need to know. The Information reports both labs drew on Astra's outputs to train new models released this month. That includes Moonshot's Kimi K3.1 Pro. Reuters says they relied heavily on those outputs. Distillation means training a model on another model's answers. The shift follows Anthropic's crackdown. In February, Anthropic said it found industrial-scale distillation campaigns by DeepSeek, Moonshot and MiniMax, using about 24,000 fraudulent accounts to extract Claude's capabilities. In October it added new defenses, including stricter usage limits for users who keep trying to extract reasoning. Anthropic says those cut distillers' harvested data by 90-99%. With Claude harder to harvest, Astra has become the main target. Astra is OpenAI's newest model. OpenAI reports 57.9% on one benchmark, against 37.3% for GPT-5.6 Sol and 55.8% for Claude Fable 5.1. It also reports 1.9x faster task completion than GPT-5.6 Sol on Mind2Web. The timing hurts Moonshot. Artificial Analysis released Intelligence Index v5 on Oct 28, with the private share of its benchmarks raised from 45% to 60%. Kimi K3.1 Pro's score dropped from 55 to 46. The Information's report landed a couple of hours later. Analysts say the models may now share Astra's style, and some argue a lab that distills can't count as frontier. Key numbers: - Astra: 57.9% on the cited benchmark vs 37.3% for GPT-5.6 Sol - Claude Fable 5.1: 55.8% on the same benchmark - Kimi K3.1 Pro on Artificial Analysis: 55 to 46 - Private share of the AA index: 45% to 60% - Anthropic's defenses: 90-99% less harvested data - About 24,000 fraudulent accounts in the February Claude campaigns The report is based on The Information's findings, and neither lab has been quoted in the material.
Open source - MT
MTS@MTSlivePost on X ·
SITUATION DETECTED: Moonshot AI and DeepSeek both distilled OpenAI’s GPT-6 Astra for their latest models. Kimi K3.1 Pro and DeepSeek’s V4.1-Flash both benefited, per The Information.
Open source