What is Anthropic accusing Alibaba of?

Anthropic alleges that operators affiliated with Alibaba's Qwen lab ran a coordinated 'adversarial distillation' campaign against Claude, using close to 25,000 fraudulent accounts to harvest 28.8 million conversations and copy Claude's software-engineering and agentic-reasoning behavior.

How does the Alibaba campaign compare to earlier distillation attacks?

Anthropic says the Alibaba-linked operation is the largest known Claude extraction attempt, surpassing the combined ~16 million-exchange campaigns it previously attributed to DeepSeek, Moonshot AI, and MiniMax (about 24,000 accounts in total) earlier in 2026.

Which Claude model did the campaign target?

According to Anthropic's letter, the campaign specifically aimed at the most advanced and commercially valuable capabilities of the Mythos Preview model — software engineering and agentic reasoning — which Anthropic frames as the cornerstones of its current frontier system.

Has Alibaba responded to the accusations?

Alibaba has not commented publicly. CNBC, Reuters, and other outlets that obtained Anthropic's letter all report that Alibaba did not respond to requests for comment when the story became public on June 24, 2026.

What is Anthropic asking U.S. lawmakers to do?

The June 10 letter went to Senate Banking Committee Chair Tim Scott and Ranking Member Elizabeth Warren ahead of an AI hearing. Anthropic urges senators to back the U.S. government's efforts to counter such distillation campaigns, with lawmakers reportedly weighing further sanctions on Chinese firms found extracting U.S. AI outputs.

Anthropic · 2026-06-24 · major

Anthropic accuses Alibaba Qwen of largest-ever Claude distillation attack

Anthropic told U.S. senators that operators tied to Alibaba's Qwen lab ran 28.8 million Claude conversations across nearly 25,000 fake accounts between April 22 and June 5, 2026 — its largest known distillation attack on Claude.

Anthropic and Alibaba logos illustrating Anthropic's distillation-attack accusation

Anthropic tells U.S. senators that nearly 25,000 fake accounts tied to Alibaba's Qwen lab harvested 28.8M Claude conversations between April 22 and June 5, 2026.

Quick facts

Accuser	Anthropic
Accused	Operators tied to Alibaba's Qwen lab
Campaign window	April 22 – June 5, 2026
Fraudulent accounts	~25,000
Exchanges harvested	28.8 million+
Target capabilities	Software engineering + agentic reasoning in Mythos Preview
Letter recipients	Sens. Tim Scott and Elizabeth Warren (June 10, 2026)

What is it?

Anthropic's June 10 letter to U.S. Senate Banking Committee leaders, made public on June 24, 2026, accuses operators affiliated with Alibaba's Qwen research division of running the largest known 'distillation' campaign against Claude. The letter says the operation pulled outputs from Anthropic's Mythos Preview model through tens of millions of fraudulent API calls.

How does it work?

Between April 22 and June 5, 2026, the campaign used close to 25,000 fake accounts to generate more than 28.8 million Claude exchanges, concentrated on Mythos Preview's software-engineering and agentic-reasoning skills. Adversarial distillation trains a weaker model on a stronger one's responses so it can copy the stronger model's behavior without paying the training cost.

Why does it matter?

The Alibaba-linked operation dwarfs the combined ~16 million-exchange campaigns Anthropic previously attributed to DeepSeek, Moonshot AI, and MiniMax. With the letter now in front of U.S. senators drafting AI legislation, the practical signal for builders is policy fallout — potential new sanctions, export-control changes, and partner-vetting requirements around Chinese frontier labs.

Who is it for?

AI policy watchers, Anthropic API customers, teams evaluating Qwen for production use

Frequently asked questions

What is Anthropic accusing Alibaba of?: Anthropic alleges that operators affiliated with Alibaba's Qwen lab ran a coordinated 'adversarial distillation' campaign against Claude, using close to 25,000 fraudulent accounts to harvest 28.8 million conversations and copy Claude's software-engineering and agentic-reasoning behavior.
How does the Alibaba campaign compare to earlier distillation attacks?: Anthropic says the Alibaba-linked operation is the largest known Claude extraction attempt, surpassing the combined ~16 million-exchange campaigns it previously attributed to DeepSeek, Moonshot AI, and MiniMax (about 24,000 accounts in total) earlier in 2026.
Which Claude model did the campaign target?: According to Anthropic's letter, the campaign specifically aimed at the most advanced and commercially valuable capabilities of the Mythos Preview model — software engineering and agentic reasoning — which Anthropic frames as the cornerstones of its current frontier system.
Has Alibaba responded to the accusations?: Alibaba has not commented publicly. CNBC, Reuters, and other outlets that obtained Anthropic's letter all report that Alibaba did not respond to requests for comment when the story became public on June 24, 2026.
What is Anthropic asking U.S. lawmakers to do?: The June 10 letter went to Senate Banking Committee Chair Tim Scott and Ranking Member Elizabeth Warren ahead of an AI hearing. Anthropic urges senators to back the U.S. government's efforts to counter such distillation campaigns, with lawmakers reportedly weighing further sanctions on Chinese firms found extracting U.S. AI outputs.