Cyber edition model

Use the Cyber edition model with security products deployed by YongRay or connect it to your existing tools through the API. Bring gateway, application and host records into one investigation to trace the affected service, the evidence behind a finding and how new records change the assessment.

How security teams can use the model

Investigate an alert

A gateway flags unusual text in a payment API request. Use its request ID to connect the alert to application events and build a timeline. If host records are missing, keep execution status open and revisit it when the block result arrives.

Prioritise affected services

Compare a security advisory with asset ownership, installed versions and external exposure. Ask for a remediation list that names each affected service, explains its priority and identifies configuration details the owner still needs to confirm.

Review changes and retests

For a change to an access check, supply a redacted change summary, the original requirement and test results. Ask which permitted and unauthorised requests have the expected outcomes and which cases still need a retest.

Watch the payment API case →

Evaluate the model with a resolved case

Choose an incident with a known outcome. Hold back the closure notes and provide redacted records in two batches. Compare the answers before and after the decisive evidence arrives.

  1. Start with the initial alert

    Ask for established facts, a source record for each fact and the questions still open.

  2. Add the decisive record

    Add a runtime block result or host event. Check whether the model identifies the conclusion that changed and the record that changed it.

  3. Compare with the closure notes

    Check event links and evidence references against the known outcome. Record time to first token, time to completion and token usage to estimate the cost of your workload.

Reference API pricing

USD per million tokens. Enterprise pricing and integration are confirmed in your quotation.

Input
US$0.30
Cached input reads
US$0.007
Output
US$1.00

Capacity and performance

Context window
1M tokens
Maximum output
393K tokens
Architecture
Transformer

Streaming lets the application display partial replies. Figures for time to first token, generation rate (tokens/s) and benchmark scores are not listed here.

Call the API

Set CYBER_API_BASE to the supplied API base address (including /v1), CYBER_MODEL to your assigned model ID and CYBER_API_KEY to your project key. Run the example on your server.

Python 3
import json
import os
from urllib.request import Request, urlopen

base = os.environ["CYBER_API_BASE"].rstrip("/")
payload = {
    "model": os.environ["CYBER_MODEL"],
    "messages": [{"role": "user", "content": "Gateway flagged request R1. Application logs are pending. List the known fact and the missing evidence."}],
}
request = Request(
    base + "/chat/completions",
    data=json.dumps(payload).encode(),
    headers={
        "Authorization": "Bearer " + os.environ["CYBER_API_KEY"],
        "Content-Type": "application/json",
    },
)
with urlopen(request, timeout=60) as response:
    print(json.load(response)["choices"][0]["message"]["content"])
JavaScript · Node.js 18+
const { CYBER_API_BASE, CYBER_API_KEY, CYBER_MODEL } = process.env;
const base = CYBER_API_BASE.endsWith("/")
  ? CYBER_API_BASE.slice(0, -1) : CYBER_API_BASE;
const response = await fetch(base + "/chat/completions", {
  method: "POST",
  headers: {
    Authorization: "Bearer " + CYBER_API_KEY,
    "Content-Type": "application/json",
  },
  body: JSON.stringify({
    model: CYBER_MODEL,
    messages: [{ role: "user", content: "Gateway flagged request R1. Application logs are pending. List the known fact and the missing evidence." }],
  }),
});
if (!response.ok) throw new Error("HTTP " + response.status);
console.log((await response.json()).choices[0].message.content);
Enquire about model access