The vendor says the model supports vulnerability research and reproduction, exploit development, penetration testing, security auditing, long-context reasoning, native function calling, and structured outputs.
The model is available through an OpenAI-compatible API as orca/orcacyber-zero-1.5. OrcaRouter lists a 1 million-token context window and a 128,000-token maximum output. List pricing is $3.00 per million input tokens, $7.50 per million output tokens, and $0.75 per million cache-read tokens.
Access is gated to the Security Research tier for trusted researchers, red teams, and authorized testing. The page also refers to an engagement, a passkey, and acceptance of terms. It reports vendor evaluation results of 95.8% on an evaluable CVE-Bench subset, 100% on Cybench unrestricted agent execution, 76.5% on SWE-bench Pro V2, and 93.9% on HumanEval+.
The source does not establish independent benchmark validation, the model's underlying architecture or weights, the assessment protocols behind the reported figures, or the actual rate limits available to approved users.