A new phase in AI-assisted software development has begun as OpenAI unveiled GPT-5.3-Codex, a powerful model designed to handle end-to-end coding workflows. The company said the model not only advances agentic coding capabilities but also marks the first time an OpenAI system has actively contributed to building itself.
The newly launched GPT-5.3-Codex is positioned as OpenAI’s most capable agentic coding model to date. It can manage full development cycles, including researching requirements, debugging entire codebases, deploying changes, and building complex applications and video games from scratch. Notably, it also played a key role in its own development by helping teams debug training runs, manage deployments, and analyse evaluation results, a process that OpenAI said significantly accelerated development.
In an official post, the company said GPT-5.3-Codex is now available globally to users on all paid ChatGPT plans across mobile and desktop apps, the command-line interface, integrated development environment extensions, and the web. Access via OpenAI’s application programming interface is expected soon.
The model combines the coding performance of GPT-5.2-Codex with the deeper reasoning and professional knowledge of GPT-5.2 into a single system. OpenAI said the model is 25% faster, enabling it to handle long and complex projects involving research, tool usage, and multi-step execution. Users can also guide the model mid-task, request progress updates, ask questions, or suggest changes without losing context.
OpenAI shared internal benchmark results to highlight performance gains. On SWE-Bench Pro, the model achieved 56.8% accuracy, compared to 56.4% for GPT-5.2-Codex and 55.6% for GPT-5.2. On Terminal-Bench 2.0, scores jumped to 77.3% from 64.0%. In OSWorld-Verified, which tests agent performance in visual desktop environments, GPT-5.3-Codex reached 64.7%, up from 38.2%.
Beyond coding, the model supports the broader software lifecycle. It can write PRDs, edit copy, conduct user research, build slide decks, analyse spreadsheets, and monitor systems. One demo showcased a fully built racing game with maps, items, and racers created from a minimal prompt.
On safety, OpenAI classified GPT-5.3-Codex as High capability under its Preparedness Framework for cybersecurity tasks. It includes safety training, automated monitoring, trusted access controls, and threat intelligence enforcement.
Also read: Viksit Workforce for a Viksit Bharat
Do Follow: The Mainstream formerly known as CIO News LinkedIn Account | The Mainstream formerly known as CIO News Facebook | The Mainstream formerly known as CIO News Youtube | The Mainstream formerly known as CIO News Twitter
About us:
The Mainstream is a premier platform delivering the latest updates and informed perspectives across the technology business and cyber landscape. Built on research-driven, thought leadership and original intellectual property, The Mainstream also curates summits & conferences that convene decision makers to explore how technology reshapes industries and leadership. With a growing presence in India and globally across the Middle East, Africa, ASEAN, the USA, the UK and Australia, The Mainstream carries a vision to bring the latest happenings and insights to 8.2 billion people and to place technology at the centre of conversation for leaders navigating the future.
