OpenAI announced on February 6 that it has released GPT-5.3-Codex, an improved agentic coding model.
According to OpenAI, GPT-5.3 Codex merges the coding performance of GPT-5.2 Codex he new Gwith the reasoning and professional knowledge of GPT-5.2, while running 25% faster.
In benchmark testing, the model recorded industry-leading performance on SWE-Bench Pro and Terminal Bench, while also showing impressive results on OSWorld and GDPVal. These four benchmarks are OpenAI’s key metrics for assessing coding ability, agent capabilities, and real-world task performance.
GPT-5.3 Codex is the first model to play a significant role in its own development. OpenAI’s Codex team leveraged the initial version for self-learning, debugging, deployment management, and result evaluation. Paid ChatGPT subscribers can now access Codex across all supported platforms where Codex is available
The company also plans to offer secure access to integrated application programming interfaces (APIs).
This update aligns with recent partnership integrations, enabling developers to use OpenAI’s Codex within Apple’s Xcode and GitHub workflows to autonomously handle complex tasks.
