Estimates assume medium reasoning effort, no batch discounts and no prompt caching. Turn on
prompt caching and the input side typically drops 50–90% —
run your own mix in the free calculator preloaded with GPT-5.5.
About this page
Prices were last verified on August 20, 2026 against OpenAI's official pricing page. Providers reprice
frequently — see our methodology for exactly how estimates are computed,
and always confirm final rates with the provider before budgeting.
GPT-5.5 FAQ
How much does GPT-5.5 cost?
GPT-5.5 charges $5 per million input tokens and $30 per million output tokens, with cached input reads at $0.50/M. Batch mode cuts both rates by 50% for asynchronous workloads.
What is the context window of GPT-5.5?
GPT-5.5 supports a context window of 1,050,000 tokens (1.1M) — roughly 787,500 English words or about 1,575 pages of text in a single request.
Does GPT-5.5 support images, audio or reasoning?
GPT-5.5 offers vision & image understanding, audio input, hidden reasoning tokens. Current frontier. Above 200K input the long-context meter applies; cache reads at $0.50/M.
How do I estimate my GPT-5.5 bill?
A typical workload of 10K input and 2K output tokens costs about $0.1700 per call with GPT-5.5. At 1,000 requests per day that is roughly $5174.37 per month. Use the free AITokenCalculator to run your exact numbers.
