distillation

Anthropic says new defenses have reduced the volume of Claude reasoning tokens extracted through distillation attacks by 90 to 99 percent. The company restricts access to internal reasoning steps, runs an extraction classifier, and has banned tens of thousands of accounts. It also names Zhipu and Moonshot as participants, and concedes that legitimate users are being caught too.

OpenAI says it detected and shut down an adversarial distillation campaign involving more than 15,000 accounts, linked in part to people associated with Moonshot AI. Researchers report that the same reasoning-extraction attack still worked on Microsoft Azure on September 13 against every OpenAI model they tested, including GPT-6 Astra.

Anthropic's latest threat intelligence report covers December 2025 through August 2026, documenting Claude's misuse in espionage, weapons development, and nationwide surveillance operations. The report also details how seven Chinese AI labs ran covert networks—some routing their own customers' requests through Claude—to extract training data at industrial scale.

Moonshot AI has released open weights, a technical report, and supporting infrastructure for Kimi K3, a model that claims 2.5x more intelligence per unit of compute. Independent testing found notable gaps in cybersecurity and math performance compared to Western frontier models.

A joint evaluation by the UK AI Security Institute and U.S. Center for AI Standards and Innovation found Kimi K3 scores 32.2% on the ExploitBench benchmark versus 76.2% for leading U.S. models, though it beats China's GLM-5.2 at 24.4%. The gap may stem from Moonshot AI distilling Claude outputs that exclude advanced offensive cyber content.

Chinese company Moonshot AI released Kimi K3, an open source model that the company claims demonstrates frontier-level performance while trailing only Claude Fable 5 and GPT 5.6 Sol. Independent analyses from Arena.ai and Vals AI suggest the model is competitive with flagship frontier models, reigniting debate about Chinese AI capabilities and open source model development.

Anthropic announced it is limiting Mythos, its newest model, to large enterprises and critical infrastructure operators rather than releasing it publicly, claiming the model's ability to discover software security exploits poses risks. The restricted rollout strategy mirrors planned approaches by OpenAI and may serve dual purposes: managing security concerns while preventing smaller competitors from using distillation techniques to replicate frontier model capabilities.

Apple has secured full access to Google's Gemini models within its data centers and is using knowledge distillation to generate training data for smaller, on-device AI models. The approach allows Apple to create lightweight versions that replicate Gemini's reasoning patterns while running directly on Apple devices, requiring significantly less processing power.