What will your LLM actually cost?

Per-token API pricing for builders — OpenAI, Anthropic, Gemini, DeepSeek, xAI and Mistral, side by side at your real volume.

28 models · 6 providers · prices verified June 2026 · independent

Step 1 — Use case

What are you building?

Drives our token-per-request estimates — you can override everything in step 2.

Step 2 — Volume & tokens

How much will you run?

Estimated from your use case — override any number.

Requests per month 100k

Input tokens / request AUTO-ESTIMATED

Output tokens / request AUTO-ESTIMATED

Input served from cache AUTO-ESTIMATED 70%

I can run batch / async — applies batch discounts where verified

Reasoning / thinking Thinking tokens bill as output — estimate how much your task adds, or set output tokens directly.

Step 3 — Results

Monthly cost per model

Grouped by tier — comparing a frontier model with a budget one on price alone is meaningless.

WizardCost · LLMCost

LLM API cost estimate — monthly cost per model

Costs marked ≈ use our auto-estimated tokens — override them in step 2 for exact math. Cached input is billed at each model's published cache-read rate; models without published cache pricing are computed at the full input rate. Models with a published long-context tier (e.g. Gemini above 200k input) switch to those rates automatically. Reasoning effort multiplies billed output to estimate thinking tokens.

Estimates computed from each provider's published per-token pricing (verified June 2026) at your selected volume; real cost depends on your exact token usage. WizardCost is editorially independent and not affiliated with any provider. Method & sources: wizardcost.com/llm/methodology.html · generated by the free calculator at wizardcost.com/llm/