OpenAI releases GPT-5.5 with improved coding and research capabilities

Investing.com -- OpenAI announced Thursday the release of GPT-5.5, its latest AI model now available to Plus, Pro, Business, and Enterprise users through ChatGPT and Codex platforms.

The model achieved 82.7% accuracy on Terminal-Bench 2.0, which tests command-line workflows, and 58.6% on SWE-Bench Pro, which evaluates GitHub issue resolution, according to benchmark results provided by OpenAI. GPT-5.5 scored 73.1% on Expert-SWE, the company’s internal evaluation for coding tasks with an estimated 20-hour completion time.

OpenAI reports that GPT-5.5 maintains the same per-token latency as GPT-5.4 while using fewer tokens to complete similar tasks. The model was developed using NVIDIA (NASDAQ: NVDA) GB200 and GB300 NVL72 systems.

For professional applications, GPT-5.5 scored 84.9% on GDPval, which tests knowledge work across 44 occupations, and 78.7% on OSWorld-Verified, which measures computer environment operation capabilities.

The model shows improved performance in scientific research applications, achieving 25.0% on GeneBench compared to GPT-5.4’s 19.0%, and 80.5% on BixBench for bioinformatics analysis.

OpenAI rated the model’s biological and cybersecurity capabilities as "High" under its Preparedness Framework and is offering specialized access through its "Trusted Access for Cyber" program for verified security professionals.

API pricing for GPT-5.5 will be $5 per million input tokens and $30 per million output tokens, with a 1-million token context window. GPT-5.5 Pro will cost $30 per million input tokens and $180 per million output tokens.

The model underwent safety evaluations including external testing and feedback from approximately 200 early-access partners before release, according to OpenAI.

You May Also Be Interested In

• The Quest to Power AI Is Going Underground
• Canada intends to extend steel and aluminum tariff measures
• UK lawmaker Jess Asato sues Elon Musk’s xAI over deepfakes