OpenAI has started releasing GPT 5.1 Codex Max, a new version of Codex built to handle coding tasks with far better speed and accuracy.
The company shared on X that this updated model can work on its own for hours, making it more reliable for long or complex projects. While GPT 5.1 is made for conversations, research, and image creation, Codex is fully tuned for programming.
Codex runs inside the terminal, but developers can also use it on the web after linking it to GitHub. Over the past few months, Codex has improved quickly and is now competing directly with Claude Code. In our testing, Codex became stronger than Claude for complicated questions after GPT 5.1 launched, although Claude still performs better when strict adherence is needed.
OpenAI says GPT 5.1 Codex Max is faster, more capable, and uses tokens more efficiently. It can stay focused on long tasks without drifting, thanks to new compaction features. The company also says this is the first Codex model trained to work in Windows environments, and it now handles PowerShell much better, making it more useful for Windows developers.
The model outperforms GPT 5.1 Codex on real engineering tasks and shows better results on SWE Bench Verified using medium reasoning, all while using around 30 percent fewer thinking tokens.
GPT 5.1 Codex Max is available starting today through the Codex CLI, IDE extension, cloud tools, and code review features.
If this article helped you, please consider supporting our work. Every small contribution keeps Abijita.com independent and running.
