Capricorn models
Gravity hosts two demo models for builders. Capricorn Mini and Capricorn Pro are made up for this demo, and so are their specs below. What really answers in the chat: a cloud model, Google Gemini 3.5 Flash Lite, reached through OpenRouter and run by Vibe Engine, with answers capped at about 500 tokens. Vibe Engine works with whichever model you choose.
Capricorn Mini
Context: 128k (demo). Best for: chat, docs search, and pricing estimates.
Input $0.40 / M tokens · Output $1.20 / M (demo table numbers).
Capricorn Pro
Context: 200k (demo). Best for: agent tasks and longer plans.
Input $2.00 / M · Output $6.00 / M (demo table numbers).
| Model | Context | Input $/M | Output $/M |
|---|---|---|---|
| Mini | 128k | $0.40 | $1.20 |
| Pro | 200k | $2.00 | $6.00 |
Open chat to try either model with the fixed action list on this site.