OPENAI API · USAGE-AWARE

OpenAI API Cost Calculator

Estimate GPT-6 Astra, GPT-6.1 Sol, or GPT-6 Luna costs from monthly tasks when you do not know token counts—or enter measured usage for an exact official list-price calculation.

Official OpenAI pricingVerified October 11, 2026USD · First-party OpenAI API
Processing mode

No token count? Estimate a monthly cost range from task type, complexity, and uses per month. Have OpenAI usage data? Choose Enter measured usage for an exact official list-price calculation.

SIMPLE ESTIMATE

Tell us what you do

SAME TASK · OFFICIAL PRICING

Compare estimated cost ranges

These ranges use each measured model’s observed usage for the same task, size, and count. They do not compare quality or speed.

OpenAI · GPT-6 Astra$1.08–$1.09No positive savings range
OpenAI · GPT-6 Luna$0.010–$0.013Potential monthly savings: $0.20–$0.22
Claude · Opus 5.5$0.87–$0.92No positive savings range
Claude · Sonnet 5.5$0.38–$0.42No positive savings range
Claude · Haiku 5.5$0.015–$0.018Potential monthly savings: $0.20–$0.21
Google · Gemini 4 Argon$0.25–$0.26No positive savings range
Google · Gemini 3.8 Flash$0.27–$0.32No positive savings range
Google · Gemini 3.6 Flash$0.29–$0.33No positive savings range

CLEAR PRICING MATH

How OpenAI API cost is calculated

Each mutually exclusive input category stays separate, and long-context pricing is applied only to usage from requests that individually exceed 272K prompt tokens.

non-cached input + cached input + cache writes + output/reasoning + regional surcharge

Batch and Flex use 50% of Standard rates. Fast uses 2× Standard. Regional processing adds 10% only for a verified supported combination.

01

Processing modes

Standard is interactive. Batch is asynchronous with up to a 24-hour completion window. Flex trades latency and availability for lower cost. Fast charges a speed premium.

02

Prompt caching

Cache writes cost 1.25× ordinary input. Cache reads cost 10% for Astra and Luna, but 5% for GPT-6.1 Sol. Savings use measured cache tokens only.

03

272K request threshold

A request over 272K prompt tokens uses 2× input and cache rates plus 1.5× output for the whole request. Monthly aggregate tokens do not determine this tier.

QUICK EXAMPLES

Example monthly workloads

GPT-6.1 Sol · Standard

1M input + 200k output/reasoning

$4.00 / month
GPT-6.1 Sol · Batch

1M input + 200k output/reasoning

$2.00 / month
GPT-6 Luna · Cached

1M cached input + 200k output/reasoning

$0.11 / month

For the most accurate result: copy actual token counts from the OpenAI response usage object. Reasoning tokens are billed as output tokens. Free or promotional credits, tools, taxes, negotiated pricing, and AWS Bedrock are not treated as public fixed rates.