OpenAI launches GPT-5.5 with better coding and research skills

OpenAI has released GPT-5.5, its latest artificial intelligence model, on Thursday. The new model is designed to perform better in coding, computer tasks, and scientific research. This launch comes just six weeks after the company introduced GPT-5.4.

Core Features and Performance

GPT-5.5 can handle complex tasks with less human help. It can plan tasks, use tools, and even check its own work without needing step-by-step instructions. According to Greg Brockman, the most impressive thing about this model is how much it can do with very little guidance.

In performance tests, GPT-5.5 scored 82.7% on Terminal-Bench 2.0, which checks how well AI models use command line tools. This is higher than Claude Opus 4.7, which scored 69.4%. On the Intelligence Index, GPT-5.5 scored 60, slightly ahead of Claude Opus 4.7 and Gemini 3.1 Pro Preview, both at 57.

However, GPT-5.5 did not lead in all tests. On SWE-Bench Pro, Claude Opus 4.7 performed better with 64.3% compared to GPT-5.5’s 58.6%. In multilingual question-answering, GPT-5.5 scored 83.2%, while Opus 4.7 and Gemini 3.1 Pro scored higher.

Pricing and Technology

The standard version of GPT-5.5 costs $5 per million input tokens and $30 per million output tokens. This is double the price of GPT-5.4. The Pro version is more expensive, costing $30 per million input tokens and $180 per million output tokens.

GPT-5.5 is the first fully retrained base model since GPT-4.5. It has a new structure, improved training data, and better focus on task-based work. It also supports a large 1 million token context window, allowing it to handle longer inputs.

As OpenAI puts it:
“GPT 5.5 reaches state-of-the-art performance across multiple benchmarks that reflect this kind of work. On GDPval⁠⁠, which tests agents’ abilities to produce well-specified knowledge work across 44 occupations, GPT 5.5 scores 84.9%. On OSWorld-Verified, which measures whether a model can operate real computer environments on its own, it reaches 78.7%. And on Tau2-bench Telecom, which tests complex customer-service workflows, it reaches 98.0% without prompt tuning. GPT 5.5 also performs strongly across other knowledge work benchmarks: 60.0% on FinanceAgent, 88.5% on internal investment-banking modeling tasks, and 54.1% on OfficeQA Pro.”

Rollout and Competition

GPT-5.5 is now available to Plus, Pro, Business, and Enterprise users in ChatGPT and Codex. OpenAI said it currently has 4 million active Codex users and 9 million paying business users. ChatGPT has over 900 million weekly users and more than 50 million subscribers.

The quick release of GPT-5.5 shows how fast AI companies are moving. With GPT-5.4 launched just weeks ago, this update highlights strong competition among leading AI developers trying to attract business customers.