The latest upgrade to the Codex ecosystem has arrived as OpenAI introduces GPT 5.1 Codex Max, a new agentic coding model designed for large scale, continuous and deeply technical software engineering work. The model is now available in Codex across the CLI, IDE extension, cloud and code review, with API access coming soon.
This new version is built on an enhanced foundational reasoning model trained across software engineering, math and research tasks. It is faster, more intelligent and more efficient with tokens, delivering stronger performance at every stage of development. GPT 5.1 Codex Max is the first model trained to operate across multiple context windows using a process called compaction. This allows it to work coherently over millions of tokens and enables project level refactoring, complex debugging and extended agent loops that run for hours.
GPT 5.1 Codex Max has been trained on real world engineering tasks such as pull request creation, code review, frontend development and technical Q and A. It outperforms earlier models in key coding evaluations and is the first Codex model trained to operate in Windows environments. Its training also includes tasks designed to improve how it collaborates within the Codex CLI.
The model shows significant gains in token efficiency through improved reasoning. On SWE bench Verified, it achieves better results than GPT 5.1 Codex while using thirty percent fewer thinking tokens at medium reasoning effort. A new extra high reasoning effort option is also available for non latency sensitive work, although medium remains the recommended default.
These efficiency gains translate to savings for developers. The model can generate high quality frontend designs at much lower cost than previous versions. Compaction also helps avoid context window limits, allowing the model to continue long tasks by pruning history and keeping only the essential context. Codex automatically compacts a session when it nears its limit and continues until the task is finished. Internal tests show that GPT 5.1 Codex Max can work independently for hours and has even completed tasks that ran for more than a day.
The model also performs better in long horizon coding and cybersecurity evaluations. Although it does not reach high capability under the preparedness framework, it is the most capable cybersecurity model OpenAI has deployed so far. The organisation is preparing additional safeguards as these abilities continue to evolve and has already disrupted attempts to misuse its models. Codex runs by default in a secure sandbox with restricted file access and no network connection, and developers are encouraged to maintain these protections.
As Codex grows more capable, reviewing its work before deployment remains essential. It provides detailed logs, tool call references and test outputs, but it should be considered an additional reviewer rather than a replacement for human checks.
GPT 5.1 Codex Max is available today for users on ChatGPT Plus, Pro, Business, Edu and Enterprise plans. It will replace GPT 5.1 Codex as the default model across Codex environments. While GPT 5.1 remains a general purpose model, Codex Max is recommended specifically for agentic coding tasks.
This launch marks another step toward more reliable long running AI systems. OpenAI reports that ninety five percent of its engineers now use Codex each week and are producing about seventy percent more pull requests since adopting it. With the new model and updates across the CLI, IDE, cloud and review tools, developer productivity is set for another major boost.
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 formerly known as CIO News is a premier platform dedicated to delivering latest news, updates, and insights from the tech industry. With its strong foundation of intellectual property and thought leadership, the platform is well-positioned to stay ahead of the curve and lead conversations about how technology shapes our world. From its early days as CIO News to its rebranding as The Mainstream on November 28, 2024, it has been expanding its global reach, targeting key markets in the Middle East & Africa, ASEAN, the USA, and the UK. The Mainstream is a vision to put technology at the center of every conversation, inspiring professionals and organizations to embrace the future of tech.
