LLM Cost Calculator: Compare GPT, Claude and Gemini Prices
Free LLM cost calculator. Enter your requests and tokens to compare monthly API costs for GPT, Claude, Gemini, Grok, DeepSeek and Mistral.
[+] IN THIS BRIEFING
- 01. How the LLM cost calculator works
- 02. Current LLM API prices per million tokens
- 03. How to use the result
- 04. Frequently asked questions
Use this free LLM cost calculator to see what your AI workload would cost on 15 current models from six providers. Enter your monthly requests and tokens, and the models re-rank instantly. Prices come from each provider’s official pricing page and were last checked on 5 November 2026. Your numbers stay in your browser.
How the LLM cost calculator works
API providers charge per token, with separate rates for input and output. Output usually costs several times more than input, so a small change in answer length can move the bill more than a large change in prompt length. The calculator applies this formula to every model:
monthly cost = requests x ((uncached input x input rate) + (cached input x cached rate) + (output x output rate)) / 1,000,000
For example, 100,000 requests a month with 1,500 input tokens, 300 output tokens and 30% of input cached comes to 180 million tokens. At today’s list prices that costs $25.95 a month on GPT-6 Luna, $514 on Claude Sonnet 5.5 and $2,595 on GPT-6 Astra.
Current LLM API prices per million tokens
| Model | Provider | Input | Cached input | Output |
|---|---|---|---|---|
| GPT-6 Astra | OpenAI | $10 | $1 | $50 |
| GPT-6.1 Sol | OpenAI | $2 | $0.1 | $10 |
| GPT-6.1 Terra | OpenAI | $1 | $0.05 | $5 |
| GPT-6 Luna | OpenAI | $0.1 | $0.01 | $0.5 |
| Claude Fable 5.1 | Anthropic | $10 | $0.25 | $50 |
| Claude Opus 5.5 | Anthropic | $4 | $0.2 | $20 |
| Claude Sonnet 5.5 | Anthropic | $2 | $0.1 | $10 |
| Claude Haiku 5.5 | Anthropic | $0.1 | $0.01 | $0.5 |
| Gemini 4 Argon ($4 / $20 from 1 Dec 2026) | $2 | $0.1 | $10 | |
| Gemini 3.8 Flash ($1.5 / $7.5 from 1 Jan 2027) | $0.75 | $0.075 | $3.75 | |
| Gemini 3.5 Flash-Lite | $0.3 | $0.03 | $2.5 | |
| Grok 4.7 | xAI | $2 | $0.5 | $6 |
| DeepSeek V4 Pro | DeepSeek | $1.32 | $0.044 | $3.96 |
| DeepSeek V4.1 Flash | DeepSeek | $0.3 | $0.006 | $1.2 |
| Mistral Large | Mistral | $0.5 | n/a | $1.5 |
Standard rates in USD, checked 5 November 2026. DeepSeek is shown at its peak rate, and off-peak is half. Mistral says caching can cut input cost by up to 90% but publishes no exact rate. Claude Haiku 5.5 and Grok 4.7 charge more for very long prompts, and the calculator applies those rates. Official sources: OpenAI · Anthropic · Google · xAI · DeepSeek · Mistral.
How to use the result
- Measure real tokens. Take your averages from production logs, because estimates often miss system prompts and retrieved documents.
- Test caching. Put a realistic cache share in the calculator, as repeated system prompts and documents are where the largest input savings usually sit. Cache-write fees are not included.
- Compare quality too. The lowest price only counts if the model passes your evaluation. A cheap model that needs retries or human review can cost more overall.
- Plan for price changes. Some current rates are promotional. Google has already announced higher Gemini 4 Argon and Gemini 3.8 Flash prices, and the calculator switches to them on their start dates.
To go deeper, read our token cost analysis for LLMs and how to reduce inference spend. Then use the AI ROI calculator to test whether the spend pays back, or the LLM VRAM calculator to compare it with self-hosting.
Frequently asked questions
How do I calculate LLM API cost?
Multiply your monthly input tokens by the input rate and your monthly output tokens by the output rate, using rates per million tokens, then add the two. If part of your input is cached, charge that share at the cached rate.
Is this LLM cost calculator free?
Yes. It is free, needs no sign-up, and runs in your browser, so the numbers you enter are not sent anywhere.
Which LLM API is the cheapest?
It depends on your mix of input, output and caching, so the ranking can change with your workload. Enter your own numbers to see it, then confirm that the cheapest model passes your quality tests.
How up to date are the prices?
Prices were last checked on 5 November 2026 against each provider's official pricing page. Providers change prices often, so confirm on the official page before you commit a budget.
Last updated: 05 November 2026
