Discussion
Moonshot's Kimi K3 AI model performs significantly below the cyber capabilities of AI models released by US frontier labs
Kimi guardrails also failed to stop users from developing exploits
https://www.aisi.gov.uk/blog/preliminary-assessment-of-kimi-k3s-cyber-capabilities
@campuscodi
Why doesn't the US trendline show what models were used to draw it?
@DefectiveWings Because the report was compiled by US and UK govt institutes, which may not want to give their endorsement for certain models
That's my 2c
@campuscodi That's not how I read that chart. I read it as they are less than 6 months behind, which is to say not really behind at all.
Install bonfire.cafe
Get the full app experience
.png)
