Models
The models available on each Perch plan, and how Roost chooses between them.
Perch runs on a range of models rather than a single one. Roost picks between them for you on every task, and on Pro you can pin a specific model by hand. This page lists what each plan includes.
This list evolves
Starter
A curated pool of capable hosted models, selected by Roost. There are no manual pins on Starter, and it includes up to $20 of hosted usage each month.
Example models: DeepSeek V4, GLM 5, and Kimi K2.5.
Pro
Every model in the registry, selectable by hand or left to Roost. Pro adds the premium models and includes up to $150 of premium-model usage each month, plus unlimited everyday use under fair use.
Example premium models: GLM 5.2, DeepSeek V4 Pro, Kimi K2.7 Code, Kimi 2.6, MiniMax M2.5, and Nemotron Ultra.
BYO
Run Perch on your own cloud key or a local OpenAI-compatible endpoint. There are no Perch usage limits on your own inference, and hosted Perch models stay off until you add a key.
At a glance
| Starter | Pro | BYO | |
|---|---|---|---|
| Model selection | Curated pool, automatic | Every model, automatic or pinned | Your own model |
| Premium models | No | Yes | Your choice |
| Manual model pin | No | Yes | n/a |
| Included usage | Up to $20 / month | Up to $150 / month + unlimited everyday use | Your own inference |
| Vision for text-only models | Yes | Yes | Yes |
For prices and plans, see Pricing. For how the routing works, see Perch Roost.
Perch Roost
Roost is Perch's routing engine. It picks the right model for each task, races across providers for speed and reliability, and lets text-only models work with images.
Security and data
How Perch handles your files and your data, what stays on your device, what is processed in the cloud, and the providers involved.