LLM API Cost Calculator (2026)
Compare what the Claude, OpenAI GPT and Google Gemini APIs will actually cost for your workload. Set your request volume and token sizes below — the table updates live, sorted cheapest first.
Start from a workload
≈ 60M input + 12M output tokens per month (30,000 requests). Assumes 30-day month.
| Model | $ / 1k requests | Monthly cost | vs cheapest |
|---|---|---|---|
Gemini 2.5 Flash-LiteCheapest Google · fast | $0.360 | $10.80 | — |
GPT-4o mini OpenAI · fast | $0.540 | $16.20 | 1.5× |
GPT-3.5 Turbo OpenAI · fast | $1.60 | $48.00 | 4.4× |
Gemini 2.5 Flash Google · fast | $1.60 | $48.00 | 4.4× |
Claude Haiku 4.5 Anthropic · fast | $4.00 | $120.00 | 11.1× |
Gemini 2.5 Pro Google · balanced | $6.50 | $195.00 | 18.1× |
Gemini 3.1 Pro Preview Google · frontier | $8.80 | $264.00 | 24.4× |
GPT-4o OpenAI · balanced | $9.00 | $270.00 | 25.0× |
Claude Sonnet 4.6 Anthropic · balanced | $12.00 | $360.00 | 33.3× |
Claude Opus 4.6 Anthropic · frontier | $20.00 | $600.00 | 55.6× |
GPT-4 Turbo OpenAI · frontier | $32.00 | $960.00 | 88.9× |
Standard pay-as-you-go rates (per 1M tokens), updated 2026-07-15. Prompt caching, batch discounts, and context-window tiers can change real costs substantially — verify against each provider’s pricing page before committing.
Dig into each provider’s pricing
Automating workflows? Try the automation cost calculator. Comparing tools instead of models? See the dev-tool pricing tracker — 28 tools, weekly-verified, with a change history.