Tokens and pricing
One balance for GPT, Claude, GLM and DeepSeek. Charges depend on input and output tokens, cached input and the selected model multiplier.
5.709 RUB per 1M balance tokens
A 600 RUB top-up provides 105.097 million balance tokens. The rate depends on the top-up amount. Minimum: 25 RUB; maximum: 50,000 RUB. Your exact credit is shown before payment.
Model multipliers
| Model | Multiplier | Priority | Input / output RUB per 1M | Cached input RUB per 1M |
|---|---|---|---|---|
GPT 6 Astragpt-6-astra | ×11.25 | ×22.5 | 64.226 | 25.69 |
GPT 6 Lunagpt-6-luna | ×0.8 | ×1.6 | 4.567 | 1.827 |
GPT 6.1 Solgpt-6.1-sol | ×4 | ×8 | 22.836 | 9.134 |
GPT 6 Solgpt-6-sol | ×3.8 | ×7.6 | 21.694 | 8.678 |
GPT 5.5gpt-5.5 | ×4.5 | ×9 | 25.69 | 10.276 |
GPT 5.6 Terragpt-5.6-terra | ×1.5 | ×3 | 8.563 | 3.425 |
GPT 5.6 Solgpt-5.6-sol | ×4.5 | ×9 | 25.69 | 10.276 |
GPT 5.6 Lunagpt-5.6-luna | ×0.9 | ×1.8 | 5.138 | 2.055 |
Z.ai GLM-5.2glm-5.2 | ×1.5 | Not applicable | 8.563 | 3.425 |
Z.ai GLM-5.3glm-5.3 | ×1.5 | Not applicable | 8.563 | 3.425 |
Z.ai GLM-5.3 Flashglm-5.3-flash | ×1 | Not applicable | 5.709 | 2.284 |
DeepSeek V4.1 Flashdeepseek-flash | ×0.3 | Not applicable | 1.713 | 0.685 |
DeepSeek V4 Prodeepseek-v4-pro | ×0.2 | Not applicable | 1.142 | 0.457 |
Claude Opus 5.5claude-opus-5-5 | ×20 | Not applicable | 114.18 | 45.672 |
Claude Fable 5.1claude-fable-5-1 | ×20 | Not applicable | 114.18 | 45.672 |
Qwen Text Embedding V4text-embedding-v4 | ×1 | Not applicable | 5.709 | 2.284 |
Claude Opus 5claude-opus-5 | ×12 | Not applicable | 68.508 | 27.403 |
Claude Opus 4.8claude-opus-4-8 | ×12 | Not applicable | 68.508 | 27.403 |
Claude Opus 4.7claude-opus-4-7 | ×10 | Not applicable | 57.09 | 22.836 |
Claude Opus 4.6claude-opus-4-6 | ×10 | Not applicable | 57.09 | 22.836 |
Claude Sonnet 5.5claude-sonnet-5-5 | ×11 | Not applicable | 62.799 | 25.12 |
Claude Haiku 4.5claude-haiku-4-5-20251001 | ×4 | Not applicable | 22.836 | 9.134 |
Claude Haiku 5.5claude-haiku-5-5 | ×9 | Not applicable | 51.381 | 20.552 |
Prices are shown for a 600 RUB top-up, excluding personal multipliers and bonuses. Cached input is charged at 40% of the regular input rate. Priority multiplies the GPT rate by 2.
Standard charge: (uncached input + output + cached input × 40%) × model multiplier. Personal terms apply separately where available.
Image generation and audio transcription use separate multipliers shown in your dashboard. Connect to the API.