Pricing
Pay per token, with every request processed only in the UK and zero data retention.
Qwen3.6 35B A3B
Model Details- Input
- £0.08/ 1M tokens
- Cached input
- £0.02/ 1M tokens
- Output
- £0.75/ 1M tokens
Paying
Prepaid balance
Top up by card with any amount from £5 to £1,000. Each request's cost comes off your balance.
Automatic top-ups
Your balance tops up from a saved card when it runs low.
Price changes
We email you 30 days before a price changes.
| Billing address | Tax | Automatic top-ups |
|---|---|---|
| UK, Channel Islands, Isle of Man | No VAT | Available |
| Other countries we sell to | Local tax added at checkout | Top up by hand |
Definitions
- Input
- Uncached prompt tokens, images included.In
usage:prompt_tokensminuscached_tokens - Cached input
- The start of a prompt that matches a recent request from the same API key, in blocks of 2,096 tokens, so shorter prompts aren't cached. Name a user or session to keep each one's cache separate.In
usage:prompt_tokens_details.cached_tokens - Output
- Every token the model writes, reasoning included.In
usage:completion_tokens, with reasoning incompletion_tokens_details.reasoning_tokens
Example
- Prompt
- 22,145 tokens, 62.5% cached
- Output
- 356 tokens
- Cost
- £0.00121, or £1.21 per 1,000 requests