CAISI report says Kimi K3 leads open-weight models but trails U.S. frontier systems
mattsheehan88 · x · 2026-07-24
- A Commerce/CAISI-NIST report says the newest Chinese model is still behind top U.S. frontier models, but is the strongest open-weight model.
- The attached charts compare cyber capability and exploit-development performance across models, including Kimi K3 and GLM-5.2.
- In the exploit-development benchmark, top U.S. models score about 76.2%, while Kimi K3 is around 32.2% and GLM-5.2 around 24.4%.
- A second chart places Kimi K3 near the top of the Chinese model cluster on overall cyber capability, though still below the U.S. frontier band.
More from Models
- 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
- In cyber, intention matters little: Chintala on model info leaks, and what users should do — brhydon · 2026-10-27
- Why don't more LLMs ship 10M-token context? A researcher asks about attention tradeoffs — agarwl_ · 2026-10-27
- Ortet's first health eval: 34 open-weight models, 50 tasks, generic rankings fall short — thegautamkamath · 2026-10-27