LLM API Cost Calculator
Estimate your monthly API spend across 10 models. Pick a use case preset — chatbot, document summarization, coding assistant — or enter raw token counts. See per-request, daily, monthly, and yearly projections for any number of users. Export the full comparison to CSV.
~500 input + 300 output per message · 100 msg/user/day
Token estimates are typical averages derived from public research on each use case. Switch to Token Mode to input your own values.
| # | Model | Provider | Input $/M | Output $/M | / Req | / Day | / Month | / Year |
|---|---|---|---|---|---|---|---|---|
| 1 | Meta | $0.00 | $0.00 | $0.00 | $0.00 | $0.00 | $0.00 | |
| 2 | DeepSeek | $0.27 | $1.10 | $0.000465 | $0.0465 | $1.40 | $16.97 | |
| 3 | Alibaba | $0.40 | $1.60 | $0.000680 | $0.0680 | $2.04 | $24.82 | |
| 4 | $1.25 | $10.00 | $0.00362 | $0.3625 | $10.88 | $132.31 | ||
| 5 | Mistral AI | $3.00 | $9.00 | $0.00420 | $0.4200 | $12.60 | $153.30 | |
| 6 | OpenAI | $2.50 | $10.00 | $0.00425 | $0.4250 | $12.75 | $155.13 | |
| 7 | Anthropic | $3.00 | $15.00 | $0.00600 | $0.6000 | $18.00 | $219.00 | |
| 8 | Perplexity AI | $3.00 | $15.00 | $0.00600 | $0.6000 | $18.00 | $219.00 | |
| 9 | xAI | $3.00 | $15.00 | $0.00600 | $0.6000 | $18.00 | $219.00 | |
| 10 | OpenAI | $10.00 | $40.00 | $0.0170 | $1.70 | $51.00 | $620.50 |
Prices reflect published API list rates as of May 2026. Free/open-weight models show $0.00 where applicable. Actual costs may vary with volume discounts. Methodology →
How costs are calculated
API costs are based on the published list prices from each provider as of May 2026. Input and output tokens are billed separately at different rates. The formula is: cost = (inputTokens/1M × inputRate) + (outputTokens/1M × outputRate). Monthly projections assume 30 days; yearly projections assume 365 days.
Use case token estimates
Use Case Mode uses typical token counts derived from published benchmarks and developer surveys. A customer support bot averages ~500 input + 300 output tokens per message. A coding assistant averages ~800 input + 600 output per query. These are approximations — switch to Token Mode for precise custom inputs.