← Models

GPT-5.5 Pro

OpenAI · GPT-5

Still-available premium GPT-5.5 API route for pinned high-compute reasoning and long-running professional workflows.

Part of GPT family · Other versions: GPT-5.6 Luna , GPT-5.6 Terra , GPT-5.6 Sol , GPT-5.5 , GPT-Rosalind , GPT-5.4 mini , GPT-5.4 nano , GPT-5.4 Pro , GPT-5.4 , GPT-5.3 , GPT-5.3-Codex , GPT-5.2-Codex , GPT-5.2 Pro , GPT-5.2 , GPT-5-Codex , GPT-5 mini , GPT-5 nano , GPT-5 , GPT-4.1 , GPT-4o mini Transcribe , GPT-4o mini TTS , GPT-4o Transcribe , Computer Use Preview , GPT-4o mini , GPT-4o
Type
language
Context
1M tokens
Max Output
128K tokens
Status
current
Input
$30/1M tok
Output
$180/1M tok
API Access
Yes
License
proprietary
frontier reasoning high-accuracy coding analysis tool-use research long-context
Released April 2026 · Updated July 10, 2026

Overview

Freshness note: Model capabilities, limits, and availability can change quickly. This profile is a point-in-time snapshot last verified on July 10, 2026.

GPT-5.5 Pro is the higher-compute member of OpenAI’s GPT-5.5 release. It launched on April 23, 2026 and remains a documented, non-deprecated API model under gpt-5.5-pro. The GPT-5.6 launch changes the default escalation path, however: OpenAI now documents Pro as reasoning.mode: "pro" on the selected GPT-5.6 model rather than publishing a separate GPT-5.6 Pro API slug.

That makes GPT-5.5 Pro primarily relevant for pinned integrations, regression comparisons, or teams that need its known behavior. New quality-first deployments should compare GPT-5.6 Sol in standard and Pro modes before choosing this older 30/30/180 route.

Capabilities

GPT-5.5 Pro is designed for harder questions and more demanding work than standard GPT-5.5 Thinking. OpenAI’s launch materials emphasize business, legal, education, data science, scientific research, manuscript critique, technical argument review, and complex analysis over large files or web context.

The main value is deeper test-time compute over the same underlying GPT-5.5 model. In practice, that should matter most when the answer needs stronger structure, better checking, and fewer missing constraints rather than just faster output.

Technical Details

OpenAI’s system card says GPT-5.5 Pro uses the same underlying model as GPT-5.5 with a setting that uses parallel test-time compute. The system card generally treats GPT-5.5 safety results as strong proxies for GPT-5.5 Pro, while still evaluating Pro separately where extra compute could materially change risk.

OpenAI’s API docs list gpt-5.5-pro with a 1,050,000-token context window, 128,000-token maximum output, text and image input, and text output. The model is designed for long-running requests; OpenAI specifically points API users toward background mode to avoid timeouts when a request may take several minutes.

The Pro API surface is not identical to regular GPT-5.5. Current docs list function calling and structured outputs as supported, streaming and fine-tuning as not supported, and the Responses tool surface excludes apply patch, skills, computer use, and tool search while still supporting web search, file search, image generation, code interpreter, hosted shell, and MCP.

Pricing & Access

GPT-5.5 Pro is available in ChatGPT for Pro, Business, and Enterprise users, and OpenAI’s API docs now list gpt-5.5-pro for developer use through the Responses API, including Batch requests.

Published API pricing for gpt-5.5-pro is:

  • Input: $30.00 per 1M tokens
  • Output: $180.00 per 1M tokens

OpenAI’s current model page does not list a cached-input discount for GPT-5.5 Pro. Regional-processing endpoints carry a 10% uplift.

Best Use Cases

Use GPT-5.5 Pro for second-pass review, executive-ready memos, difficult architecture analysis, legal or policy synthesis with human review, scientific research support, and complex data-science reasoning. It is especially appropriate after a cheaper model has already narrowed the problem.

Avoid using it as a default chat model for routine drafting, simple coding, extraction, or support triage. The economics only make sense when the answer quality delta matters.

Comparisons

  • GPT-5.6 Sol (OpenAI): Current flagship with a 5/5/30 standard route and configurable Pro mode instead of a separate Pro model slug.
  • GPT-5.5 (OpenAI): Better default for hard agentic work where Pro-level spend is not justified.
  • GPT-5.4 Pro (OpenAI): Previous OpenAI Pro escalation tier, still available for existing GPT-5.4 integrations but superseded in product positioning by GPT-5.5 Pro.
  • Claude Opus 4.7 (Anthropic): Natural cross-vendor comparison for premium reasoning, coding, and long-horizon work.
  • Gemini 3.1 Pro Preview (Google): Preview alternative when Google-native multimodal and long-context integrations matter.