Editorial illustration for OrcaCyber Zero 1.5 Model Boasts Top-Tier Cyber Scores at USD 3 Per 1M Tokens
OrcaCyber Zero 1.5 Hits 100% Cybench at $3/1M Tokens
OrcaCyber Zero 1.5 Model Boasts Top-Tier Cyber Scores at USD 3 Per 1M Tokens
OrcaRouter put a price tag on frontier-level vulnerability hunting this week: $3.00 per million input tokens, $7.50 per million output tokens, for a model it says clears 100% on Cybench. OrcaCyber Zero 1.5 is the successor to Zero 1.0, which launched September 17, 2026, and it's built specifically for authorized security work rather than general-purpose chat. OrcaRouter lists the use cases plainly:
OrcaRouter has released OrcaCyber Zero 1.5, a model for authorized vulnerability research. The model is the successor to OrcaCyber Zero 1.0, which shipped on September 17, 2026. OrcaCyber Zero 1.5 is a post-trained Orca model for vulnerability reproduction, exploit development and penetration testing.
Why this matters
The price tag is the pitch here: $3/$7.50 per million tokens for scores OrcaRouter says match frontier cyber models. That's worth watching if it holds up, but the 24-task CVE-Bench sample is thin enough that we'd treat the benchmark as a teaser, not a verdict. Self-reported numbers from the vendor shipping the model are not independent validation, and security tooling is exactly the category where inflated scores cause real damage, whether that's missed vulnerabilities or false confidence in a pen-testing pipeline.
The 1M-token context and native function calling matter more concretely: those are the features that let a model actually trace a codebase or chain exploit steps instead of just answering questions about them. For security teams and researchers, the useful move is to run your own evaluation set before trusting the marketing framing, especially the "fewer reports, more validated fixes" line, which is a business outcome, not a benchmark result. Zero 1.0 shipped roughly two months before this release, so OrcaRouter is iterating fast.
Fast iteration cuts both ways when the claims outrun the independent testing.
Common Questions Answered
What are the pricing and performance claims for OrcaCyber Zero 1.5?
OrcaCyber Zero 1.5 is priced at $3.00 per million input tokens and $7.50 per million output tokens, with OrcaRouter claiming the model achieves a perfect 100% score on Cybench. The vendor positions this as frontier-level pricing for vulnerability hunting capabilities comparable to other advanced cyber models.
What specific security use cases is OrcaCyber Zero 1.5 designed for?
OrcaCyber Zero 1.5 is purpose-built for authorized security work including vulnerability reproduction, exploit development, and penetration testing. Unlike general-purpose chat models, it is specifically optimized for these specialized cybersecurity applications rather than broad conversational use.
How does OrcaCyber Zero 1.5 differ from its predecessor Zero 1.0?
OrcaCyber Zero 1.5 is the successor to OrcaCyber Zero 1.0, which launched on September 17, 2026. While both are post-trained Orca models for security work, the 1.5 version represents an updated iteration with improved capabilities for vulnerability research and authorized security testing.
Why should the Cybench benchmark scores for OrcaCyber Zero 1.5 be viewed with caution?
The 24-task CVE-Bench sample used to validate the model's performance is considered too limited to serve as definitive proof of capability. Additionally, self-reported benchmark numbers from the vendor shipping the model lack independent third-party validation, which is particularly concerning in security tooling where inflated scores can lead to real damage like missed vulnerabilities.
Further Reading
- Cybench: A Framework for Benchmarking Cybersecurity Capability of Language Models - Cybench
- Cybench: AI Cybersecurity Benchmark - Emergent Mind
- How AI Finds Zero-Days: The 2026 Guide - o-mega
- OrcaCyber Zero 1.0 API — Pricing & Benchmarks - OrcaRouter