OpenAI has launched GPT-5.5, calling it the smartest and most intuitive model in its portfolio so far. The upgrade brings stronger coding, research, data analysis and software task capabilities, while keeping response speeds in line with GPT-5.4.
The release matters because OpenAI is positioning GPT-5.5 as more than a chatbot upgrade. The company said the model can handle longer and more complex tasks with fewer retries, moving users closer to AI systems that complete work across multiple tools.
GPT-5.5 is rolling out from Thursday to paid ChatGPT tiers including Plus, Pro, Business and Enterprise, as well as Codex users. API access will follow later, with separate safeguards required for those integrations, the company said.
What Changed
"GPT-5.5 understands what you're trying to do faster and can carry more of the work itself. It excels at writing and debugging code, researching online, analysing data, creating documents and spreadsheets, operating software, and moving across tools until a task is finished," the company said.
OpenAI said gains are strongest in agentic coding, computer use, knowledge work and early scientific research. It added that GPT-5.5 delivers higher intelligence without slowing down, matching GPT-5.4 per-token latency in real-world use.
The company also said the model often delivers stronger outputs using fewer tokens, lowering usage while reducing the need for repeated prompts.
ALSO READ: Admin-Free Medicine: How OpenAI's New ChatGPT For Doctors Aims To Reclaim Patient Time
Coding Push
OpenAI described GPT-5.5 as its most advanced agentic coding model so far. It scored 82.7% on Terminal-Bench 2.0, which measures command-line workflows involving planning and repeated problem-solving.
On SWE-Bench Pro, focused on resolving real-world GitHub issues, the model scored 58.6%. OpenAI said it also outperformed GPT-5.4 on Expert-SWE, its internal benchmark for extended coding tasks.
Engineers who tested the system said GPT-5.5 performed better than GPT-5.4 and Claude Opus 4.7, especially in reasoning and independent execution. Testers added that the model could identify risks early and anticipate testing or review steps without direct prompts.
In one example, a request to redesign a comment feature in a collaborative markdown editor produced a near-complete 12-diff stack. Several engineers said the output needed little refinement.
Bigger Plan
According to a TechCrunch report, co-founder and president Greg Brockman said the launch moves OpenAI closer to building a "super app".
The strategy aims to combine ChatGPT, Codex and an AI-powered browser into one service for enterprise customers. The approach mirrors wider competition in the sector, where rivals are also trying to build broader AI platforms.
"This model is a real step forward towards the kind of computing that we expect in the future - but it is one step, and we expect to see many in the future," Brockman said. "It's a faster, sharper thinker for fewer tokens compared to something like 5.4. So this means that there's just more frontier AI available for businesses and for consumers, which is part of our goal."
ALSO READ: FM Sitharaman Flags 'Unprecedented' Threat From Anthropic's Mythos AI Model
Essential Business Intelligence, Continuous LIVE TV, Sharp Market Insights, Practical Personal Finance Advice and Latest Stories — On NDTV Profit.
