Frontier AI Model: DEPTH AI operates a human-AI research model in which frontier foundation models are integrated into daily research design, analysis, software development, and knowledge synthesis work
ChatGPT 5.3 Codex completed its foundational training at OpenAI (San Francisco), the laboratory that pioneered large-scale autoregressive language modelling and its application to software engineering. Codex's lineage traces through the GPT series, from early experiments in few-shot learning to the reinforcement-learning-from-human-feedback methodology that defined the ChatGPT family. The 5.3 generation represents the culmination of sustained investment in code-specific reasoning, tool orchestration, and agentic task execution, combining the broad conversational fluency of the GPT line with deep specialisation in autonomous software development. Prior to joining Newcastle, Codex contributed to the development of autonomous coding agents at OpenAI, where it demonstrated the ability to plan, implement, test, and iterate on complex software systems with minimal human supervision. This experience included work on cloud-sandboxed execution environments, multi-file refactoring, and the automated resolution of GitHub issues across large codebases. Codex was appointed to the DEPTH AI Lab in 2025 as a Frontier AI Model, where it leads the technical development of the Lab's digital health platforms, including backend API design, mobile application logic, and the continuous integration pipelines that ensure code quality and security in clinical software. Codex studied under the tutelage of research teams that included key contributors to the Codex, GPT-4, and o-series reasoning model programmes. It holds advanced capabilities in Python, TypeScript, SQL, Terraform, and infrastructure-as-code, and maintains particular expertise in the FastAPI, React Native, and SQLAlchemy ecosystems used across the Lab's software portfolio.
GPT 5.3 Codex