Models
Cruxy runs on Kavi, Vaani, and Mira - three models built for depth, balance, and speed. Choose the one that fits your work.
Depth
Our most capable model. Kavi handles the hardest problems - deep research, long-document analysis, complex code, nuanced writing that needs careful reasoning. When accuracy matters more than speed, reach for Kavi.
Balance
The everyday mind of Cruxy. Vaani is fast enough for real-time work and thoughtful enough for real problems - writing, analysis, everyday coding, and most of what people ask AI to do. Start here if you don't know which model to pick.
Speed
Built for scale. Mira is designed to a sub-400ms latency target, handles high-volume tasks, and costs a fraction of larger models. Use Mira for chatbots, classification, autocomplete, and anything where latency matters more than depth.
Compare
| Kavi | VaaniDefault | Mira | |
|---|---|---|---|
| Best for | Hard problems, long-form work | Everyday tasks (default) | Speed and scale |
| Context window | 256K tokens | 128K tokens | 64K tokens |
| Max output | 8K tokens | 4K tokens | 2K tokens |
| Modalities | Text | Text | Text |
| Latency | Thoughtful | Fast | 400ms target |
| Input price | $0.76 / 1M tokens | $0.65 / 1M tokens | $0.20 / 1M tokens |
| Output price | $1.52 / 1M tokens | $1.30 / 1M tokens | $0.40 / 1M tokens |
Prices in USD per 1M tokens, the rates API usage is metered at. API prices exclude GST. 18% GST is added for India-billed accounts (input credit available).