GPT-5.5 token & cost calculator

GPT-5.5 is OpenAI's current flagship — the top of the GPT-5 line and the priciest of OpenAI's general models at $5 input / $30 output per million tokens. It pairs frontier reasoning with a roughly 1,050,000-token context window, but with a twist most OpenAI models don't have: a long-context surcharge that doubles the input rate (to $10/$45) once a request crosses 272,000 input tokens.

That surcharge changes the calculus for big-context work. GPT-5.5 is priced for hard reasoning, not for cheap long-context throughput — for the latter, GPT-4.1's flat $2/$8 across the same million-token window is usually the better buy. Reach for GPT-5.5 when reasoning quality is the binding constraint and reserve the huge-context requests for when you actually need both.

Expected response (output tokens)

Prompt

Client-side. Never uploaded.

0 / 1,000,000 charactersContext window: 1,050,000 tokens

Or start with an example

Total estimated cost

$0.031GPT-5.5

Tokensexact

Input cost

$0.00

Output cost (est.)

$0.031

@ 1,024 response tokens

Context used

of 1,050,000

Verified 2026-07-06 · exact

Saved scenariosnone yet

Saved on this browser only — never uploaded. Up to 10 scenarios.

Tip: save a scenario when you have a prompt + model + response length you might revisit. Useful for sizing features before committing to a vendor.

Verify privacysince this page loaded — updates live

Prompt uploads0Always 0 — by design

Outgoing requests0Analytics + page assets only — no prompt content

Cookies on this origin0Vercel Analytics + Clarity may set first-party cookies

localStorage keys0Theme preference + saved scenarios live here

Server endpoints1/api/og only — accepts title + subtitle, never prompt text

Inspect

Open DevTools → Network. Type into the calculator. No request bodies should contain your prompt text.

Pricing

GPT-5.5 is tier-priced on input size: $5/$30 at or below 272K input tokens, $10/$45 above it. The calculator applies the correct tier automatically.

Tier	Input $/M	Output $/M
≤ 272,000 input tokens	$5	$30
All input	$10	$45
Context window	1,050,000 tokens

Verified against developers.openai.com on 2026-07-06.

Worked examples

The 200K example bills at the standard $5/$30; the 400K and near-full examples cross 272K and re-rate to $10/$45. Paste your own prompt above to price it exactly.

Scenario	Input	Output	Cost
Short chat turn A typical reasoning turn with a small system prompt.	800	400	$0.016
200K input (standard tier) A large prompt still under the 272K surcharge threshold.	200,000	4,000	$1.12
400K input (surcharge tier) Crosses 272K — the whole request re-rates to $10/$45.	400,000	6,000	$4.27
Near-full context Deep in the surcharge tier, close to the ~1.05M ceiling.	900,000	8,000	$9.36

The step between the 200K and 400K rows is steep because two things move at once: more tokens and a higher per-token rate. For sustained work above 272K, model the surcharge tier as your real rate — and price the same prompt on GPT-4.1 before committing, since its flat rate often wins at that scale.

How is this counted?

GPT-5.5 uses the o200k_base tokenizer, shared with GPT-5.4, GPT-4.1, and GPT-4o. We count via gpt-tokenizer (MIT) — the canonical OpenAI vocab, exact match, calibration factor 1.0. The count of your pasted text is exact; note that the live API adds a few fixed formatting tokens per message, so invoiced input runs slightly higher than raw content. Inputs over 50,000 characters tokenize in a Web Worker.

FAQ

Is the token count exact?

Yes. GPT-5.5 uses OpenAI's canonical o200k_base tokenizer, which gpt-tokenizer ships verbatim, so the count of your text is exact — no approximation or calibration factor. One caveat: the Chat/Responses API adds a few fixed formatting tokens per message on top of your content, so your invoiced input runs a handful of tokens higher per message than the raw text count.

Why does the price change on very large prompts?

GPT-5.5 has a long-context surcharge. At or below 272,000 input tokens it's $5 input / $30 output per million; above 272,000 the whole request switches to $10 input / $45 output — 2× input and 1.5× output. The calculator applies the right tier automatically based on your input size.

How does GPT-5.5 compare to GPT-4.1 on cost?

GPT-4.1 is much cheaper ($2/$8, flat) and also carries a million-token window, so for pure long-context throughput where you don't need GPT-5.5's reasoning, GPT-4.1 usually wins on price. GPT-5.5's premium buys frontier reasoning quality; reach for it when that's the binding constraint, not when you just need to fit a big input.

What is the context window?

Roughly 1,050,000 tokens of input, with a maximum output of 128,000 tokens. The long-context surcharge starts at 272,000 input tokens — about a quarter of the window — so most large requests are billed at the upper tier.

Should I use GPT-5.5 or GPT-5.4?

GPT-5.4 is the mid tier at $2.50/$15 — half the price of GPT-5.5's $5/$30 — and shares the same tokenizer. Use GPT-5.4 for most production work and step up to GPT-5.5 only for the hardest requests where the extra reasoning earns its cost. The plain GPT-5 (original) still runs at $1.25/$10 but is deprecated, with shutdown scheduled for December 2026.

Compare against every other model

To see this exact prompt scored against every supported model, sorted by total cost, paste it into the home calculator and toggle Compare across all models. The comparison is most revealing on prompts above 272K tokens, where GPT-5.5's surcharge separates it from flat-priced long-context models.

The natural comparison set: GPT-5.4 (the mid tier at half the price), GPT-4.1 (the flat-priced million-token option that undercuts GPT-5.5 on pure long-context cost), and Claude 4.8 Opus (the cross-vendor frontier-reasoning comparison).