Perimattic
Free calculator

LLM API pricing calculator: Claude, OpenAI, Gemini

Pick models from Anthropic, OpenAI and Google, enter your monthly requests and token counts, and see the monthly cost of the same workload on each model side by side, with prompt caching and batch discounts.

Free to usePrices checked November 17, 2026Claude, OpenAI, Gemini

Last updated October 22, 2026

US one-dollar bills scattered across a flat surface
Overview

How much does the Claude API cost compared with OpenAI and Gemini?

On November 17, 2026, Claude Opus 5.5 cost $4 per 1M input tokens and $20 per 1M output tokens, Claude Sonnet 5.5 $2 and $10, and Claude Haiku 5.5 $0.10 and $0.50. OpenAI’s gpt-6.1-sol cost $2 and $10, and Google’s Gemini 4 Argon $2 and $10, rising to $4 and $20 on December 1. Batch APIs halve these prices.

Calculator

What will your workload cost on each model?

Tick the models you want to compare, then enter your monthly requests and the average input and output tokens per request. If you do not know your token counts yet, measure a few real prompts with the token calculator. Add the share of input your provider's prompt cache would serve and the share of requests that can wait for the batch API. The result lists every model you picked, cheapest first.

The starting workload is an example — replace it with your own. Prices are the providers' list prices, checked on November 17, 2026.

Anthropic models
OpenAI models
Google models

API calls your application makes in a typical month.

System prompt, context and the user message together.

The model’s reply, including any reasoning tokens billed as output.

Share of input tokens billed at the cached input price. Use 0 if you do not cache prompts.

Share of requests that can wait for asynchronous batch processing. Use 0 for none.

Sign up once to see your results. It's free, and the results open here as soon as the form is sent.

Show results

List prices in US dollars, before tax, volume discounts and cache-write or cache-storage charges. Models tokenize text differently, so the same prompt is not the same number of tokens on every model. The calculator runs in your browser: nothing you type is sent to a server. Seeing the result needs a one-time free sign-up, and only the sign-up form is sent.

How it works

How is the monthly cost worked out?

Each request is priced from the provider's list prices per 1M tokens. Uncached input, cached input and output are billed separately, and requests sent through the batch API use the batch prices. The example uses the calculator's starting workload on Claude Sonnet 5.5; the workload is an example, not a benchmark.

Starting workload: 1,000,000 requests a month, 1,500 input and 400 output tokens per request, 30% of input cached, 20% of requests batched.
StepFormulaExample (Claude Sonnet 5.5)
1. Split the inputInput tokens × (1 − cached %) and input tokens × cached %1,500 × 70% = 1,050 uncached; 450 cached
2. Standard cost per request(Uncached × input price + cached × cached price + output × output price) ÷ 1M(1,050 × $2.00 + 450 × $0.10 + 400 × $10.00) ÷ 1M = $0.006145
3. Batch cost per requestThe same, at batch prices(1,050 × $1.00 + 450 × $0.05 + 400 × $5.00) ÷ 1M = $0.0030725
4. Monthly costStandard requests × step 2 + batch requests × step 3800,000 × $0.006145 + 200,000 × $0.0030725 = $5,531
5. Cost per 1,000 requestsMonthly cost ÷ requests × 1,000$5.53
6. Yearly costMonthly cost × 12$66,366

The calculator leaves out cache-write charges, Google's hourly cache storage, long-prompt surcharges, regional processing uplifts and tax. If those matter for your workload, read the provider pages in our LLM API pricing guides. To see what caching your own repeated requests or routing simple ones to a cheaper model could save, use the LLM gateway savings calculator.

Price table

What do Claude, OpenAI and Gemini models cost per 1M tokens?

These are the prices the calculator uses, in US dollars per 1M tokens, read from each provider's own pricing page. Providers change prices often, so confirm on the source page before you commit to a budget.

Anthropic API prices

Prices checked on November 17, 2026. Cached input is the cache hit (read) price. Writing to the cache costs 1.25× input (5-minute cache) or 2× input (1-hour cache). The Batch API takes 50% off input and output, and Anthropic says batch and caching discounts combine, so batch cached input is shown at half the cache hit price.
ModelInputCached inputOutputBatch input / output
Claude Fable 5.2$8.00$0.20$40.00$4.00 / $20.00
Claude Fable 5.1$10.00$0.25$50.00$5.00 / $25.00
Claude Opus 5.5$4.00$0.20$20.00$2.00 / $10.00
Claude Opus 5$5.00$0.50$25.00$2.50 / $12.50
Claude Sonnet 5.5$2.00$0.10$10.00$1.00 / $5.00
Claude Sonnet 5$2.00$0.20$10.00$1.00 / $5.00
Claude Sonnet 4.6$3.00$0.30$15.00$1.50 / $7.50
Claude Haiku 5.5Prompts up to 100,000 tokens. Longer prompts: $0.50 input, $0.05 cached input, $2.50 output.$0.10$0.01$0.50$0.05 / $0.25
Claude Haiku 4.5$1.00$0.10$5.00$0.50 / $2.50

Source: Anthropic pricing page.

OpenAI API prices

Prices checked on November 17, 2026. Standard tier, prompts of 272K input tokens or fewer; longer prompts cost more. Batch prices are OpenAI’s published Batch tier.
ModelInputCached inputOutputBatch input / output
gpt-6-astra$10.00$1.00$50.00$5.00 / $25.00
gpt-6.1-sol$2.00$0.10$10.00$1.00 / $5.00
gpt-6.1-terra$1.00$0.05$5.00$0.50 / $2.50
gpt-6-sol$2.00$0.20$10.00$1.00 / $5.00
gpt-6-luna$0.10$0.01$0.50$0.05 / $0.25
gpt-5.6-solOpenAI’s page says promotional pricing for GPT-5.6 Sol is available until November 30, 2026.$4.00$0.40$20.00$2.00 / $10.00
gpt-5.6-terra$2.00$0.20$12.00$1.00 / $6.00
gpt-5.6-luna$0.20$0.02$1.20$0.10 / $0.60
gpt-5.5$5.00$0.50$30.00$2.50 / $15.00
gpt-5.4$2.50$0.25$15.00$1.25 / $7.50
gpt-5.4-mini$0.75$0.075$4.50$0.375 / $2.25
gpt-5.4-nano$0.20$0.02$1.25$0.10 / $0.625

Source: OpenAI API pricing page.

Google API prices

Prices checked on November 17, 2026. Paid tier, Standard, text input. Cached input is the context caching price; Google also charges for cache storage by the hour on some models, which is not included.
ModelInputCached inputOutputBatch input / output
Gemini 4 CarbonIntroductory price. From December 1, 2026: $4.00 input, $0.20 cached input, $20.00 output.$2.00$0.10$10.00$1.00 / $5.00
Gemini 4 ArgonIntroductory price. From December 1, 2026: $4.00 input, $0.20 cached input, $20.00 output.$2.00$0.10$10.00$1.00 / $5.00
Gemini 4 Hydrogen$1.50$0.075$7.50$0.75 / $3.75
Gemini 3.8 FlashIntroductory price through December 31, 2026. From January 1, 2027: $1.50 input, $0.15 cached input, $7.50 output.$0.75$0.075$3.75$0.375 / $1.875
Gemini 3.6 FlashIntroductory price through December 31, 2026. From January 1, 2027: $1.50 input, $0.15 cached input, $7.50 output.$0.75$0.075$3.75$0.375 / $1.875
Gemini 3.5 Flash-Lite$0.30$0.03$2.50$0.15 / $1.25
Gemini 3.1 Flash-Lite$0.25$0.025$1.50$0.125 / $0.75
Gemini 2.5 Flash-Lite$0.10$0.01$0.40$0.05 / $0.20

Source: Gemini API pricing page.

Choosing a model

How do you compare LLM prices fairly?

A lower price per token does not always mean a lower bill. Before you switch models on price alone:

  1. Use your own token counts. On the models listed here, output costs four to just over eight times as much as input, so a model that writes long answers can cost more than its input price suggests.
  2. Count tokens on each model. Tokenizers differ, so the same prompt can be a different number of tokens on Claude, GPT and Gemini models.
  3. Check quality on a test set. A cheaper model that fails more often costs more in retries and review time.
  4. Check scheduled changes.Google's page lists price rises for Gemini 4 Argon and Carbon on December 1, 2026 and for Gemini 3.8 and 3.6 Flash on January 1, 2027.
  5. Count the costs around the model. Logging, tracing and hosting add to the bill; estimate them with the LLM observability cost calculator and the AI infrastructure cost planner.

If a fine-tuned smaller model could replace a large one for a narrow task, our LLM fine-tuning services team can test that, and our AI development services cover building the application around the model. For costs of AI features built on Perimattic AI Suite, try the AI Suite LLM cost calculator. If you are comparing AI coding tools rather than APIs, use the Claude Code vs Cursor cost calculator.

Close-up of colour-highlighted code on a dark monitor
LLM Gateway

How does Perimattic LLM Gateway help with multi-provider costs?

Perimattic LLM Gateway sits between your application and model providers. It offers:

Get started: Sign up and our team will get in touch, or read the LLM Gateway overview.

Small plant growing out of a pile of coins
FAQs

LLM API pricing questions

How is Claude API pricing calculated?

Anthropic bills each request by tokens: input tokens at the model’s input price, output tokens at its output price, both quoted per million tokens. Input read from the prompt cache is billed at the lower cache hit price, and writing to the cache costs 1.25 times input for the 5-minute cache. The Batch API takes 50% off input and output.

Is there an OpenAI API pricing calculator for cached input?

Yes. Set the share of input tokens read from the cache and this calculator bills that share at the cached input price on OpenAI’s pricing page, for example $0.10 per 1M tokens on gpt-6.1-sol against $2.00 for uncached input. OpenAI lists separate cache-write prices on some GPT-6 and GPT-5.6 models, which are not included here.

Which LLM API is cheapest?

It depends on the workload. Among the models listed here on November 17, 2026, Claude Haiku 5.5, gpt-6-luna and Gemini 2.5 Flash-Lite shared the lowest input price, $0.10 per 1M tokens, and Gemini 2.5 Flash-Lite had the lowest output price, $0.40. A cheaper model only saves money if it handles your task well, so test quality on your own prompts.

How much does the Gemini API cost?

On November 17, 2026, Google listed Gemini 4 Argon and Gemini 4 Carbon at $2 per 1M input tokens and $10 per 1M output tokens on the paid tier, doubling to $4 and $20 on December 1, 2026. Gemini 3.5 Flash-Lite cost $0.30 and $2.50, and Gemini 2.5 Flash-Lite $0.10 and $0.40.

Why does the same prompt cost different amounts on different models?

Two reasons. Each model has its own price per token, and each provider’s tokenizer splits the same text into a different number of tokens. Anthropic says its newer tokenizer, used from Claude 4.7, produces about 30% more tokens for the same text. Count tokens with each provider’s own tool before you compare costs closely.

Do these prices include tax, long prompts or reasoning tokens?

They are list prices before tax and volume discounts. OpenAI charges more for prompts over 272K tokens and Claude Haiku 5.5 for prompts over 100,000 tokens; this calculator uses the standard rates. Reasoning or thinking tokens are billed as output tokens, so include them in your average output tokens per request.

LLM Gateway

Want one API for every model, with costs under control?

Sign up to tell us which models and providers you use, and our team will get in touch to set up the LLM Gateway for your stack.
  • Free consultation
  • No commitment
  • Tailored to your team