I picked a premium model, but the results aren't any better. Why?

Why this happens

A better model cannot help if the agents never get to finish their reading. A subscription OAuth token has rate limits, so the provider throttles the scan while it runs. The agents then cover less of the codebase. Cyberlop says it plainly: "Premium models only pay off with an API key."

An amber box on the AI Provider tab repeats the warning when you pick the OAuth option. If you chose a premium model for any tier and stayed on that token, the results are unlikely to beat the built-in defaults, and the extra cost buys you nothing.

How to fix it

1. Open Settings, then AI Provider.
2. Check Provider type. If it shows the OAuth token option, swap it.
3. Choose the API key option and paste your key into Replace credential. Press Save credential and then Verify.
4. Scroll to AI Models. The three tiers are Deep analysis, Recon & exploitation and Lightweight tasks.
5. If you do not need a specific model, leave a tier on its default. Press Save models.
6. Run the scan again and compare the lines of code analysed.

Still not working?

• A premium model changes how deeply the agents reason. It does not widen the scope of the profile, so check focus and avoid rules too.
• Read how the model tiers work before changing all three.
• To see the mix of included and own-key spend, open how AI usage is measured.

Need a second opinion? Reach us at /contact.

Related questions

• Why did my scan cover less of my codebase than expected?
• Should I use an API key or a subscription OAuth token for my AI provider?
• Why did an amber warning appear when I chose the OAuth token option?
• How do I pick which AI model each tier uses?
• How do I go back to the default models?

Full tutorial

AI provider settings: included allowance, your own credential and models

Step-by-step, with annotated screenshots.

Updated 8 October 2026 · Back to the help centre