Same three models, available via API. Pay only for what you use, priced 30% under comparable models from US labs.
Models
| Kavi | VaaniDefault | Mira | |
|---|---|---|---|
| Best for | Hard problems | Everyday tasks (default) | Speed and scale |
| Context window | 250K tokens | 200K tokens | 128K tokens |
| Max output | 16K tokens | 8K tokens | 4K tokens |
| Modalities | Text, vision | Text, vision | Text |
| Input price (INR) | ₹850 / 1M | ₹170 / 1M | ₹13 / 1M |
| Output price (INR) | ₹4,250 / 1M | ₹850 / 1M | ₹68 / 1M |
| Input price (USD) | $10 / 1M | $2 / 1M | $0.15 / 1M |
| Output price (USD) | $50 / 1M | $10 / 1M | $0.80 / 1M |
Prices in INR or USD, your choice. GST applies for Indian customers. Volume discounts available above 100M tokens/month.
How API pricing works
You pay only for the tokens you actually send and receive. No setup fees, no minimum monthly spend, no charge for failed requests.
Input is what you send Cruxy - your prompt, context, documents. Output is what Cruxy generates back. Output costs more because it requires inference.
At 100M tokens/month, automatic 15% discount. At 1B tokens/month, contact sales for custom rates.
Try before you scale
Every new API account gets ₹500 in credits - enough to test all three models thoroughly. No credit card required to start.
Common questions