Chat, browser Agent, and ordinary API keys share one model's limits. SacredAI and Heretic limits do not pool.
Chat plans
Free, Start, Core, Pro, and Max are Chat plans.
A plan sets separate SacredAI and Heretic windows. Hosted Hermes is not included.
Chat, browser Agent, and ordinary API
Free → Start → Core → Pro → Max
Prices are in roubles per month. Each model has its own windows.
| Capability | Free₽0per month | Start₽990per month | Core₽1,990per month | Pro₽4,990per month | Max₽9,990per month |
|---|---|---|---|---|---|
| SacredAI · limits | 10,000 / 5h50,000 / 7 days | 100,000 / 5h500,000 / 7 days | 250,000 / 5h1,500,000 / 7 days | 500,000 / 5h3,500,000 / 7 days | 1,000,000 / 5h7,000,000 / 7 days |
| Heretic · limits | 10 lifetime requests | 100,000 / 5h500,000 / 7 days | 250,000 / 5h1,500,000 / 7 days | 500,000 / 5h3,500,000 / 7 days | 1,000,000 / 5h7,000,000 / 7 days |
| Concurrent streams | 1 | 1 | 2 | 3 | 4 |
| Images per month | 5 | 100 | 300 | 1000 | 2500 |
The free 10,000 / 5h and 50,000 / 7d windows renew. Free API uses the same limit. A paid plan gives SacredAI the same numbers as the plan row, separate from Heretic.
The 10 lifetime requests are shared by Chat, browser Agent, and the Public API. After that a paid plan's separate limits apply.
Not included in Pro or Max. Hosted agents have separate Lite, Pro, and Ultra plans and their own per-agent, per-model limit.
One limit unit equals ceil(output + uncached input × 0.02 + reported cached input × 0.001) per request. This is not a per-token rouble price.
Direct line
Found a bug? Show us.
Tell us what happened and what you expected. The page address and browser details help us reproduce it faster.