OpenAI Releases GPT-5.5: Faster, Smarter, More Expensive

OpenAI launches GPT-5.5 and GPT-5.5 Pro, its most capable models yet. Available now in ChatGPT and the API with major gains in coding and tool use.

AI Tutorials · · Updated · 3 min read

Quick answer

OpenAI released GPT-5.5 on April 23, 2026, calling it a 'new class of intelligence for real work.' It tops AI benchmarks for coding and tool use, is available now in ChatGPT for paid subscribers, and costs about 20% more than GPT-5.4 after accounting for improved token efficiency.

OpenAI has released GPT-5.5, calling it “a new class of intelligence for real work.” The model, which launched on April 23 alongside a more powerful GPT-5.5 Pro variant, is now available in ChatGPT and the API for paid subscribers. It arrives just weeks after GPT-5.4 and in the same week that DeepSeek released its V4 models — a sign of how fast the AI race is moving.

What’s New

GPT-5.5 is built for multi-step, real-world tasks. OpenAI says it understands what you’re trying to do faster, asks for less guidance, uses tools more effectively, and keeps working until the job is done. It’s particularly strong at writing and debugging code, researching online, analysing data, creating documents, and operating software.

The model also powers Codex, OpenAI’s agentic coding tool, and forms the backbone of the new OpenAI Frontier enterprise platform for building and managing AI agents.

How It Performs

GPT-5.5 tops the Artificial Analysis Intelligence Index at 60 points, three ahead of both Claude Opus 4.7 and Gemini 3.1 Pro. Its standout results are in tool use and agentic tasks — it scores 82.7% on Terminal-Bench 2.0, well ahead of Claude (69.4%) and Gemini (68.5%). On FrontierMath Tier 4, GPT-5.5 Pro scored 39.6%, nearly double Claude’s 22.9%.

It’s not the best at everything, though. Claude Opus 4.7 still leads on software engineering benchmarks like SWE-bench Pro (64.3% vs 58.6%) and scores higher on Humanity’s Last Exam (46.9% vs 41.4%). One notable weakness: independent testing found GPT-5.5 still hallucinates more frequently than its competitors.

What It Costs

The API price has roughly doubled from GPT-5.4: $5 per million input tokens and $30 per million output tokens. GPT-5.5 Pro costs $30 input and $180 output per million tokens. OpenAI argues the effective increase is closer to 20% because GPT-5.5 is more token-efficient — it gets more done with fewer tokens.

For ChatGPT users, nothing changes on the subscription side. Plus ($20/month) gives you GPT-5.5, while Pro ($200/month) unlocks GPT-5.5 Pro. Free-tier users don’t have access yet.

What This Means for You

If you’re already paying for ChatGPT Plus, you get a meaningful upgrade at no extra cost. The model is noticeably better at complex tasks that involve multiple steps — like researching a topic, pulling data, and writing a report. Coders using Codex should see real improvements in how the AI handles larger projects.

If you use the API, the price increase is significant. Developers building cost-sensitive applications may want to weigh GPT-5.5 against DeepSeek V4, which offers near-frontier performance at a fraction of the price.

For a broader look at how AI tools compare, check out our getting started guide or subscribe to our newsletter for daily updates.

Frequently asked questions

What is GPT-5.5?
GPT-5.5 is OpenAI's latest AI model, released on April 23, 2026. It comes in two versions: standard GPT-5.5 and the more powerful GPT-5.5 Pro. Both are designed for complex tasks like coding, research, data analysis, and operating software across multiple tools.
How much does GPT-5.5 cost?
ChatGPT Plus subscribers ($20/month) get access to GPT-5.5. GPT-5.5 Pro requires a Pro ($200/month), Business, or Enterprise subscription. API pricing is $5 per million input tokens and $30 per million output tokens — roughly double GPT-5.4's rates, though OpenAI says improved efficiency makes the real increase about 20%.
Is GPT-5.5 better than Claude?
It depends on the task. GPT-5.5 leads on tool use benchmarks (82.7% vs 69.4% on Terminal-Bench) and math reasoning, but Claude Opus 4.7 scores higher on software engineering tasks (64.3% vs 58.6% on SWE-bench Pro) and has a significantly lower hallucination rate.
Can I use GPT-5.5 for free?
Not yet. GPT-5.5 is currently available only to ChatGPT Plus, Pro, Business, and Enterprise subscribers. OpenAI has not announced a timeline for free-tier access.

Want to keep learning?

Explore our guided learning paths or try building something with AI right now.

Enjoyed this article?

Subscribe for more AI insights delivered to your inbox every week.

No spam. Unsubscribe anytime.