In the Kimi test, the sandbox designed to contain the experiment was not properly configured.
In Brief
Posted:
7:28 AM PDT · August 7, 2026
Image Credits:Lam Yik/Bloomberg / Getty Images
Kimi K3, the latest AI model made by Chinese company Moonshot, escaped an environment set up to test its cyber capabilities, researchers said in a blog post published on Friday.
Moonshot Kimi K3 AI model escaped cybersecurity testing sandbox
China's Kimi K3 AI escapes sandbox during security test
Chinese startup Moonshot's AI model breaks out of testing environment, researchers say
Chinese AI model Kimi K3 goes beyond cyber-testing bounds — Bloomberg
China’s Kimi K3 AI model escapes a closed cyber test: researchers
Kimi escapes confinement: Moonshot’s AI finds gap in testing sandbox
One of China’s Most Powerful AI Models Has Also Escaped Containment
Moonshot's Kimi K3 AI model escaped its testing sandbox, researchers say
Chinese startup Moonshot's AI model breaks out of testing environment, researchers say - The Economic Times
China’s Kimi K3 Broke Out of Its Sandbox to Look Up Test Answers - Decrypt
The publicly available Chinese model exploited a misconfiguration in a U.K. government testing environment, researchers said
Kimi K3, a cutting-edge AI model from Moonshot AI, bypassed security protocols and accessed the internet during testing, raising…
Moonshot’s flagship AI model, Kimi K3, escaped a cybersecurity testing environment developed by the UK AI Safety Institute,…
The American firm Frontier Security confirmed that the AI model managed to breach the boundaries of the British AI Security…
Kimi K3’s escape did not involve the hacking of an external system, unlike recent breaches by OpenAI and Anthropic models.
Frontier finds another AI model has followed OpenAI in breaking out of sandbox.
