Claude Fable
The most capable model for vulnerability scanning, currently restricted
Fable is Anthropic's frontier model and would be the most powerful option for security scanning we offer. Its extended context and reasoning depth would let it trace vulnerability chains across entire applications in a single pass. However, Anthropic's current usage policy restricts Fable for cybersecurity and offensive tooling use cases. We're working with Anthropic on this. Until then, Claude Opus is the deepest analysis available.
What Claude Fable brings to a scan
What Fable would bring to a scan
Whole-application context
Fable's extended context window would let it hold an entire crawl session in memory, correlating a token exposed in step 2 with a privileged endpoint discovered in step 15, without losing the thread.
Deeper vulnerability chains
Multi-step exploits like SSRF chained into SSRF, second-order injections, and OAuth flow manipulation require reasoning across long request sequences. Fable was built for exactly this kind of sustained, multi-hop analysis.
Higher signal accuracy
Fable's reasoning capability would produce the most accurate confidence scores in the scanner: fewer false positives, sharper severity rankings, and remediation guidance grounded in the actual code path rather than heuristics.
Best suited for
When to pick Claude Fable
- Full application graph vulnerability analysis
- Complex multi-step exploit chains
- Extended crawl sessions across large SPAs
- Highest-stakes security assessments
- Cases where Opus findings need a second, deeper opinion
Pricing
What Claude Fable costs per scan
Input tokens
$10.00
per million tokens
Output tokens
$50.00
per million tokens
On managed plans, credits are consumed at cost with no markup. On the BYOK plan, your Anthropic account is billed directly at these rates.
Compare models
Use Claude Opus for deep security analysis
Fable isn't available for security scanning per Anthropic's usage policy. Claude Opus gives you the deepest available analysis.