Overview
GPT-5.5 is OpenAI's current flagship model, announced on April 23, 2026 and made available in the API on April 24, 2026. OpenAI positions it as 'a new class of intelligence for coding and professional work,' built to be a faster, sharper thinker that uses fewer tokens than the previous GPT-5.4 generation while delivering stronger results.
The model accepts text and image input and produces text output. It has a 1,050,000-token context window and can return up to 128,000 output tokens, with a knowledge cutoff of December 1, 2025. It supports streaming, function calling, structured outputs, and a broad set of hosted tools including web search, file search, image generation, code interpreter, hosted shell, computer use, and MCP.
GPT-5.5 is designed for complex agentic tasks: multi-step, tool-using workflows where the model must plan and execute over long horizons. OpenAI highlights improvements in recovering from errors mid-task, making efficient tool calls, maintaining coherence over long contexts, and calibration. It rolled out first to ChatGPT Plus, Pro, Business, and Enterprise tiers and inside the Codex coding assistant.
| Released | 2026-04 |
|---|---|
| License | Proprietary |
| Weights | API only |
| Parameters | Undisclosed |
| Context | 1.05M |
| Max output | 128K |
| Knowledge cutoff | 2025-12-01 |
| Modalities | Text, Vision |
| Status | Available |
Pricing
| Input | $5.00 / 1M tokens |
|---|---|
| Cached input | $0.50 / 1M tokens |
| Output | $30.00 / 1M tokens |
Strengths
- Frontier-level agentic coding, debugging, and long-horizon tool use
- Very large 1,050,000-token context window for long documents and codebases
- Improved token efficiency versus the prior GPT-5.4 generation
- Broad native tool support (web/file search, code interpreter, computer use, MCP)
- Strong scientific and technical research workflow performance
Best for
- Agentic software engineering: writing, debugging, and shipping code across tools
- Long-context document and data analysis
- Autonomous multi-step research and knowledge work
- Operating software and orchestrating tool-using agents
How to access
| Provider | Model ID |
|---|---|
| OpenAI API ↗ | gpt-5.5 |
GPT (Flagship / Thinking) — every version
The full lineage of the GPT (Flagship / Thinking) line, newest first. Every version has its own page — click any to compare specs, benchmarks and pricing.
FAQ
Is GPT-5.5 OpenAI's current flagship model?
Yes. GPT-5.5, announced on April 23, 2026, is OpenAI's current flagship model as of mid-2026. OpenAI describes it as a new class of intelligence for coding and professional work, designed to be faster and more token-efficient than the prior GPT-5.4 generation while excelling at agentic, tool-using tasks.
What is GPT-5.5's context window and knowledge cutoff?
GPT-5.5 has a 1,050,000-token context window and can produce up to 128,000 output tokens per response. Its training knowledge cutoff is December 1, 2025. It accepts both text and image inputs and returns text output, making it suitable for long-document and large-codebase work.
How much does the GPT-5.5 API cost?
Per OpenAI's official model docs, GPT-5.5 costs $5.00 per 1M input tokens and $30.00 per 1M output tokens, with cached input priced at $0.50 per 1M tokens. Prompts exceeding 272K input tokens are billed at higher session rates.
What is GPT-5.5 best at?
GPT-5.5 is built for complex agentic work: multi-step coding, debugging, research, and operating software across tools over long time horizons. OpenAI emphasizes better mid-task error recovery, more efficient tool calls, and improved coherence across its very large context window.