LLM Cost Calculator
Estimate model costs with explicit assumptions.
Standard · USD · paid text API
Your workload
Total input includes cache reads and writes. Output should include billable reasoning or thinking tokens.
WORKLOAD ESTIMATE
Monthly cost
—
Enter a token budget and request volume. No prompt text or API key is needed.
Ready for a workload estimate.
Standard global API text rates. Above 272,000 total input tokens, long-context rates apply to the entire request. Regional/Fast/Batch charges are not modeled.
Compare the same workload
Choose models, then calculate above. Token counts stay the same; equal text can use different counts on different tokenizers. A lower price is not a quality recommendation.
Prices can change. Estimates exclude taxes, storage/time, search, tool, media and other unmodeled charges. Actual billing is determined by the provider. Fractional currency is rounded only for display.
Enable JavaScript to calculate locally. No server submission is available.
Plan with the whole request.
Enter total input and output tokens, then requests per day and the number of days in your planning month. The default example is 10,000 daily requests with 2,000 input and 500 output tokens. All arithmetic runs locally using integer currency units.
Compare a budget, not model quality.
Comparison holds your numeric workload constant. Different tokenizers, reasoning needs and response lengths can change actual usage. A provider may reject a request beyond its limits even though arithmetic can estimate a cost.
Pricing assumptions.
How is caching handled?
Cached reads and cache writes are disjoint portions of total input. The remainder uses the ordinary input rate. Choose a supported write duration; quantities never double-count tokens. Minimum cache eligibility, hit rate and storage-duration fees are not inferred.
How do thresholds and dates work?
The total input count selects a whole-request rate band, including cached and written tokens. Dates use inclusive starts and exclusive ends. The snapshot begins October 8, 2026; earlier historical pricing is unavailable. Announced future prices have their own rules.
Is this live pricing?
No. This is reviewed static data with official source links and last-verified dates. Verification older than 30 days is flagged. The tool does not scrape providers or call model APIs.
Which fees and tiers are included?
Direct-provider paid Standard text rates only. Output includes billable reasoning/thinking tokens. Batch, priority, regional uplifts, taxes, media, grounding, storage and tool charges are excluded. Monthly volume is a projection, not an invoice.