UK AISI/CAISI finds Kimi K3 lags frontier cyber models in attack tests
ShakeelHashim · x · 2026-07-24
Kimi K3 trails frontier cyber models in UK AISI/CAISI tests
A joint UK AISI / CAISI evaluation found that Kimi K3 performs significantly below the latest frontier cyber-capable models on preliminary cyber tests.
- On exploit-development tasks, Kimi K3 ranked well below leading frontier models.
- On the simulated corporate-network attack benchmark “The Last Ones”, it averaged step 17 of 32, while the most cyber-capable U.S. models averaged 28.5 steps.
- The report says Kimi K3 still outperformed GLM-5.2 on the same evaluations.
- Its safeguards did not prevent attempts at cyber exploit development or offensive cyber operations during the tests.
More from Models
- GPT-6.1 Terra arrives, trained for fast and efficient computer use — mark_k · 2026-10-27
- Nat Lambert: narrow RL training on internal infra is a poor fit for open-weight models — mervenoyann · 2026-10-27
- Leaked Terminal-Bench 4.0 chart shows GPT-6.1 Terra leading at lowest cost, unverified — Angaisb_ · 2026-10-27
- OpenAI brings GPT-6.1 Terra to the API and Codex for fast, cheap computer use — OpenAIDevs · 2026-10-27
- GPT-6.1 Terra rolls out to all Plus, Pro, Enterprise and Edu users — OpenAI · 2026-10-27
- OpenAI launches GPT-6.1 Terra with 50% lower API prices than GPT-5.6 Terra — OpenAI · 2026-10-27