calcLLM
Estimate your LLM API costs
Compare published token rates with the cost of your workload, including repeated context and extra attempts.
Estimator mode
Naive Estimate
$31,680.00/mo
$1,056.00/day
Workload Estimate
$69,710.40/mo
$2,323.68/day
Planning estimate in USD, not an invoice or subscription quote. Standard API token rates; 30 days per month (22 workdays for coding). Use the additional reasoning-token control for billed thinking that is not replayed in conversation history; avoid double-counting it in visible output. Tool calls, search, images/audio, cache storage, taxes, regional surcharges and negotiated discounts are excluded. Retries mean extra full-cost attempts; actual failed-request billing varies. Compare low and high workload assumptions.
Claude Sonnet 5: rates reviewed 2026-09-07. Standard global API pricing. Cache scenario uses the 5-minute write price; 1-hour writes cost more. Provider pricing.
Cost Breakdown
Lowest estimates among refreshed models
based on Conversational AI defaults; cost ranking does not imply equivalent capability
- 1.OpenAI · GPT 5.6 Luna$7,564.32/moEstimate with this model →
- 2.Google · Gemini 3.5 Flash-Lite$13,422.96/moEstimate with this model →
- 3.Google · Gemini 3.8 Flash$26,141.40/moEstimate with this model →
- 4.Anthropic · Claude Haiku 4.5$34,855.20/moEstimate with this model →
- 5.Anthropic · Claude Sonnet 5$69,710.40/moEstimate with this model →