OpenAI launches GPT‑5.3-Codex as AI race heats up after Anthropic’s Claude Opus 4.6 — all you need to know
On February 5, OpenAI, led by Sam Altman, introduced its new Codex model, GPT‑5.3-Codex, which the company claims is “the most capable agency coding model to date.”
It is the first model that “makes a meaningful contribution to its own creation.” Early versions were used in training to debug, manage deployment, and analyze test results and evaluations, according to the company statement.
What can the latest Codex model do?
GPT-5.3-Codex can handle long, complex tasks that involve exploration, tooling, and multi-step execution, allowing coders to interact with the model and guide it as it works without losing context, as if collaborating with a colleague, the company statement said.
Features of GPT-5.3-Codex
OpenAI said the GPT-5.3-Codex model combines the advanced coding capabilities of GPT-5.2-Codex with the strong judgment and professional knowledge of GPT-5.2 into a single system. He also added that the new model works at least 25% faster than previous models.
what’s up
With GPT-5.3-Codex, the model has changed from an agent that used to write and review code to an agent that can perform almost every task that developers and other professionals do on the computer. Simply put, the Codex model was previously able to write and control code; but with the new upgraded version, it can perform tasks similar to what a developer or professional does. This includes researching information, organizing files, managing workflows, and operating tools.
GPT-5.3-Codex is designed to support all work throughout the software lifecycle, including debugging, deployment, monitoring, writing Product Requirements Documents (PRDs), editing copy, user research, testing, measurements, and more. It can also analyze data on pages.
availability
The latest GPT-5.3-Codex is available to all users with paid ChatGPT plans and can be used everywhere. This includes the application, Command Line Interface (CLI), Integrated Development Environment extension (IDE), and web.
Rival AI model launched by Anthropic
As the competition between the two AI companies heats up, the launch of GPT-5.3-Codex coincides with Anthropic’s new AI model, Claude Opus 4.6. The Opus 4.6 model is an improved version. Anthropic said Thursday that it builds on its predecessor with more powerful coding capabilities. Moreover, it schedules carefully, can handle long-running intermediary tasks, runs more reliably on large codebases, and can offer improved code review.
OpenAI and Anthropic
Following the launch of two competing AI models, one of the company’s leading researchers, OpenAI’s Sam Altman and Anthropic’s Sholto Douglas, appeared on a podcast. According to a report in Business ContentThe OpenAI chief said: “I think we’ll move towards a workflow where many people feel like they’re managing a team of agents. As agents evolve, they’ll continue to operate at an increasingly higher level of abstraction.”
Responding to Altman, Douglas said they compared previous models and saw significant differences. According to him, “OpenAI models were slightly better on really, really, really difficult trials on difficult problems, but the Anthropic models were much faster and so on.” He added that while OpenAI is trying to increase speed, Anthropic is focusing on making its AI models “much, much better at really, really tough problems.”
