A joint preliminary evaluation by the UK's AISI and the US' CAISI finds Kimi K3 trails leading US frontier closed weight models on cyber capability (AI Security Institute)
.png)
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- UK AISI and US CAISI conducted a joint preliminary evaluation finding that Kimi K3 trails leading US frontier closed-weight models on cyber capability.
- The evaluation specifically benchmarks frontier models against each other on cyber capability, with Kimi K3 compared directly against the top US closed-weight systems.
Why it matters: Two major national AI safety bodies — the UK's AISI and US' CAISI — are jointly benchmarking frontier models for cyber capability, putting Kimi K3 in direct comparison with leading US closed-weight systems on a dimension with clear national security implications.
.png)
.png)
