Token cost calculator
Estimate a month of LLM API usage: enter your input and output tokens and your cache hit rate, and see the cost per model on OpenScale against the list price.
DeepSeek V4.1 Flash on OpenScale, a month
$5.55
$11.10 at list price. You save $5.55 (50%).
| Tokens | Amount | Price / 1M | Cost |
|---|---|---|---|
| Input | 50M | $0.12 $0.06 | $3.00 |
| Cached input | 450M | $0.006 $0.003 | $1.35 |
| Output | 5M | $0.48 $0.24 | $1.20 |
Prices used
| Model | Input | Cached input | Output |
|---|---|---|---|
| DeepSeek V4.1 Flash | $0.12 $0.06 | $0.006 $0.003 | $0.48 $0.24 |
| GLM 5.3 Flash | $0.15 $0.12 | $0.03 $0.024 | $0.50 $0.40 |
How the estimate works
Input tokens split into cached and uncached by your cache hit rate. Each part is multiplied by its price per 1M tokens, and output tokens by the output price. The list price is the reference rate before the discount, priced on the same workload.
Your real cache hit rate is cached tokens divided by prompt tokens, from the usage field of your responses. The default here is a coding agent: almost all input, most of it repeated each turn. Reasoning tokens count as output.
Estimates only: your bill counts the exact tokens of each request. See the price comparison for an agent workload across providers, and the API reference for model limits.