AI modelsOpenAI

GPT-5.5

Decision matrix

GPT-5.5 model decision matrix
Overview

GPT-5 is OpenAI's flagship API model, priced at $1.25 per million input tokens and $10 per million output tokens, called through the OpenAI API rather than the ChatGPT consumer app. Pick it when you're building an application and want OpenAI's top-tier model directly; use ChatGPT instead if you just want to chat.

Other good fits
  • Tasks where GPT-5's extra capability over GPT-5 mini is worth roughly 5x the input price and 5x the output price.
Pros

GPT-5 is OpenAI's flagship model available through the API: the model that powers the higher tiers of ChatGPT, callable directly from your own code for building applications, not just chatting. It's the model to reach for when a task genuinely needs OpenAI's top capability rather than a faster, cheaper variant.

Pick it when…

  • You're building a product or feature that calls a model programmatically, not chatting interactively yourself.
  • The task needs OpenAI's most capable current model, not just "good enough" from a cheaper tier.
  • You want direct API access rather than routing through a third-party gateway.
Cons
Read cons

GPT-5's per-token price is meaningfully higher than GPT-mini's, so defaulting to it for every task, including simple, high-volume ones, is an easy way to overspend without a matching quality benefit. Like any model, it can still produce a confident, wrong answer; review output for anything the application will act on directly.

Pick something else when…

  • You just want to chat, not build an application. ChatGPT is the consumer product built for that.
  • The task is routine or high-volume. GPT-mini costs a fraction of GPT-5's price and may be enough.
  • You want to compare GPT-5 against models from other vendors on the same task. OpenRouter lets you do that from one integration.

Avoid if

  • The task is routine and doesn't need the flagship model's extra capability; [GPT-mini](tool page) costs a fraction of the price.
Context

Not documented yet.

Requirements

Not documented yet.

Licence & cost

Prices checked 23 September 2026 (pricing (opens in a new tab)): $1.25 per million input tokens, $10 per million output tokens, standard API pricing. There's no subscription; you pay per token through your OpenAI API account. Watch output token usage specifically, since it's priced 8x the input rate, and a verbose response costs more than a concise one for the same underlying task.

First task

Documentation-based suggestion — not run here.

Send the same non-trivial task, for example drafting a structured response to a made-up complex customer enquiry with several sub-questions, to both GPT-5 and GPT-mini, and compare the results against the roughly 5x price difference between them.

Done when you have a concrete side-by-side result and can say whether the task actually needed GPT-5's extra capability, or whether the cheaper model would have done just as well.

Links & evidence
Documentation-based · not hands-on · Sep 23, 2026

Rests on OpenAI's API pricing page, read on 2026-09-23. No API call was made for this review; context-window size was not published on the pricing page and should be confirmed on the model's own documentation page.

  • OpenAI API pricing · Sep 23, 2026 — GPT-5 at $1.25/$10 per million input/output tokens; GPT-5 mini at $0.25/$2.
  • OpenRouter pricing · Sep 23, 2026 — Comparison alternative for calling GPT-5 alongside other vendors' models.
  • ChatGPT pricing · Sep 23, 2026 — Comparison alternative for chatting directly rather than building against the API.
Further review notes & alternatives

Connection

Call GPT-5 through the standard OpenAI API with your API key, specifying the model name in your request. It follows OpenAI's usual chat/response API conventions, so existing OpenAI API client code generally needs only the model name changed to switch to it.

Alternatives and trade-offs

  • GPT-mini wins on cost for routine or high-volume tasks that don't need the flagship model's full capability.
  • Claude Opus wins as the closest cross-vendor comparison for demanding reasoning work; worth benchmarking both on your actual task.
  • ChatGPT wins when you want to use the model yourself through a chat interface rather than build against the API.

Learn more

Catalogue facts

Handle multi-step reasoning, tool use, and messy agentic tasks. Adjust reasoning effort when depth and response time need balancing. Use text and image inputs within a large context window.

OpenAI · paid · cloud

  • multi-step reasoning
  • adjustable depth
  • large context
  • image input
  • agentic workflows
  • tool integration
  • cloud inference
  • tiered pricing
  • write-code
  • automate
  • talk-think
  • draft-docs
  • data-analytics

Browse all tools