OrcaCyber Zero 1.5 reports perfect results on 39 Cybench challenges, but the figures are company-reported and the model is access-gated through OrcaRouter.
By [RuntimeWire Staff](https://runtimewire.com/author/runtimewire-staff)
· Published
Primary source: [PR Newswire](https://www.prnewswire.com/news-releases/continuum-ai-introduces-orcacyber-zero-1-5--a-frontier-cybersecurity-reasoning-model-302905663.html)
Why it matters #
OrcaCyber puts Continuum's research thesis into a sensitive, paid API product. The reported scores attract attention, but security teams will need reproducible evaluations, safe operating controls and evidence from real workflows before benchmark performance translates into trust.
Continuum AI announced OrcaCyber Zero 1.5 on October 12th, pitching the model as a way to carry security investigations from vulnerability discovery through exploit testing and patch verification.…
source & further reading
runtimewire.com — original article
Ted Haley releases Pensieve, a local search tool for coding agents
Sakana AI trains a model on its agents' collective research work
Two coding-agent tests report DeepSeek V4.1 Flash used exposed API keys