Home / Help centre / Workspace settings
Workspace settings
What are the Deep analysis, Recon & exploitation and Lightweight tasks tiers?
Cyberlop runs on three model tiers. Deep analysis covers code reading and vulnerability analysis. Recon & exploitation covers recon, proof in a real browser and the report. Lightweight tasks covers structured extraction. Each tier has defaults you can change under Settings → AI Provider → AI Models.
The short version
Cyberlop splits its AI work into three tiers. Every tier fits a kind of task, so Cyberlop spends the most capable model only where it changes the result. Leave every tier on its default if you like. You can also pick a model for each one. The choice applies to your whole organisation, not to a single scan or person, and every later scan uses it.
How it works in Cyberlop
The tiers sit under Settings → AI Provider → AI Models. Each has its own dropdown with a list of known models, plus a Custom… option that lets you type any model id, so you can adopt a new model on the day it ships.
| Tier | Used for | Size of model |
|---|---|---|
| Deep analysis | Pre-recon (reading the code) and the vulnerability analysis specialists | Largest |
| Recon & exploitation | Recon, the exploitation agents and the final report | Medium |
| Lightweight tasks | Structured extraction | Smallest |
The Deep analysis tier decides finding quality, so it gets the deepest reasoning. That is why its default is the largest model on the list. Recon and exploitation mean lots of tool use and browser work. A mid-size model suits that high-volume job well, and it keeps cost sensible. The Lightweight tier handles small extraction jobs and sits off the critical path today.
Click Save models after you change a dropdown.
What to do next
Keep the defaults unless you need a specific model, for example one your team has already approved for use with its code. A premium choice only pays off with an AI provider API key. On a rate-limited subscription token, throttling can hit a scan and it covers less code. Results may fall short of the defaults. Not worth it. Read API key or OAuth token before you change a tier.
To change a tier, follow how to pick a model for each tier. To use a model that is missing from the list, see using a custom model id.
Related questions
- How do I pick which AI model each tier uses?
- How do I use a model that isn't in the dropdown?
- How do I go back to the default models?
- Should I use an API key or a subscription OAuth token for my AI provider?
- I picked a premium model, but the results aren't any better. Why?
AI provider settings: included allowance, your own credential and models
Step-by-step, with annotated screenshots.
Updated · Back to the help centre
Still stuck?
Tell us what you see on screen and an engineer will help you sort it out.