AI Resources
LLM API cost calculator
Enter your monthly requests and the tokens in each one to see what every model in the pricing dashboard would cost.
Last verified: October 13, 2026
| Model | Provider | Input $/MTok | Output $/MTok | Monthly cost |
|---|---|---|---|---|
| Gemini 2.5 Flash-Lite | Gemini | $0.10 | $0.40 | $4.00 |
| GPT-6 Luna | OpenAI | $0.10 | $0.50 | $4.50 |
| DeepSeek V4.1-Flashcached input | DeepSeek | $0.006 | $1.20 | $6.12 |
| GPT-5.6 Luna | OpenAI | $0.20 | $1.20 | $10.00 |
| GPT-5.4 Nano | OpenAI | $0.20 | $1.25 | $10.25 |
| DeepSeek V4.1-Flash | DeepSeek | $0.30 | $1.20 | $12.00 |
| Gemini 3.1 Flash-Lite | Gemini | $0.25 | $1.50 | $12.50 |
| Gemini 3.5 Flash-Lite | Gemini | $0.30 | $2.50 | $18.50 |
| Gemini 3 Flash Preview | Gemini | $0.50 | $3.00 | $25.00 |
| Gemini 3.6 Flash | Gemini | $0.75 | $3.75 | $33.75 |
| Gemini 3.8 Flash | Gemini | $0.75 | $3.75 | $33.75 |
| GPT-5.4 Mini | OpenAI | $0.75 | $4.50 | $37.50 |
| DeepSeek V4-Pro | DeepSeek | $1.32 | $3.96 | $46.20 |
| Z.ai GLM 5.3via Mistral | Mistral | $1.40 | $4.40 | $50.00 |
| Claude Sonnet 5 | Anthropic | $2.00 | $10.00 | $90.00 |
| Claude Sonnet 5.5 | Anthropic | $2.00 | $10.00 | $90.00 |
| GPT-6 Sol | OpenAI | $2.00 | $10.00 | $90.00 |
| GPT-6.1 Sol | OpenAI | $2.00 | $10.00 | $90.00 |
| Gemini 3.1 Pro Preview | Gemini | $2.00 | $12.00 | $100.00 |
| GPT-5.6 Terra | OpenAI | $2.00 | $12.00 | $100.00 |
| Claude Opus 5.5 | Anthropic | $4.00 | $20.00 | $180.00 |
| GPT-5.6 Sol | OpenAI | $4.00 | $20.00 | $180.00 |
| Claude Opus 4.6 | Anthropic | $5.00 | $25.00 | $225.00 |
| Claude Opus 5 | Anthropic | $5.00 | $25.00 | $225.00 |
| Claude Fable 5 | Anthropic | $10.00 | $50.00 | $450.00 |
| Claude Fable 5.1 | Anthropic | $10.00 | $50.00 | $450.00 |
| GPT-6 Astra | OpenAI | $10.00 | $50.00 | $450.00 |
How the estimate works
Cost = requests × (input tokens × input price + output tokens × output price) ÷ 1,000,000, at each provider's standard rates in US dollars. Caching, batch discounts, long-context surcharges, tools and taxes are not included. A model is marked instead of priced when one request exceeds its documented context or output limit.