Google’s Jules aims to out-code Codex in battle for the AI developer stack

The coding and the growth of the Acting Coding platform led to another battlefield between technology companies.

In December, Google JULES, an independent coding factor that can repair errors as a simultaneous, as an experiment. However, through Google I/O, Google announced that JULES will be now available in Beta.

With the broader version of JULES, Google places itself as a strong competitor for an increasing number of artificial intelligence coding assistants designer to write, verify the code and fix it independently.

“It would be available to help developers to repair errors, create tests, and consult the documents that occur in the background,” Josh Woodward, Google Labs.

“People describe applications in existence,” Woodward said. “This has started as an uninterrupted coding agent with the idea that, what if you create a method by which you can assign the tasks of this agent for things that you do not want to do?”

JULES will be integrated into GitHub and Gemini 2.5 Pro uses from Google. During the general experimental version stage, developers can access the JULES for free but with the limits of use.

Inclusive and parallel

JULes operates asymmetry, allowing developers to set a task while working separately on something else. It manages the tasks inside a virtual device, shows the tasks and their causes, and even provides audio summaries.

But Gul is not the coding agent of the simultaneous tasks and the only parallel in the market, and he was not the only one who was announced in May.

Openai surprised the industry with the release of a research inspection for Codex Agent Codex, after distributing rumors that the company will buy Windsurf to start operating. Codex has started life as a coding model but has since turned into a coding agent capable of writing, fixing errors and answering the code database questions in a separate sand box.

Codex was also behind one of the first code assistants, GitHub Copilot. GitHub is announced in Microsoft Build this week, Github Copilot, who performs a lot of the same asymmetric work as Codex and JULES.

Social media is full of attention to the upcoming arms race around coding agents, even before the release of Jules and Codex in full to the public.

Yes, I think Gul outperforms the manuscript a lot. It has been tested only on a lazy claim so far, “Project Analysis and Writing Unit Tests to cover 100 %.”
Gulps plans first and creates her own tasks. The manuscript does not. This is great.
– JULES VMS has the Internet pic.twitter.com/dcgpkwinip

– Daniel Nakov (@dnak0v) May 19, 2025

@Google Anti -intelligence agent Julis presented her first contribution to a project I am working on
Reactions: I really hope there will be a way in which I can choose the files or evidence that I want to focus on artificial intelligence pic.twitter.com/z5ymaf2ERB

– Nicolas (@nicolasserna314) May 20, 2025

It seems that the coding agents that can introduce PRS are new shiny creatures. Codex from Openai, COPILOT coding agent from GitHub/Microsoft, and JULES from Google, Claude and XAI when?

Samuelsurfboard May 20, 2025

The most independent coding platforms follow this “competitive coding” growth, where software and applications are often created by claiming instead of difficult coding by humans. The entrance to the major technology companies such as Google and Openai in this arena brings more coding agents to the Vanguard of AI Arms.

More symbol that works with the same Amnesty International

Even inside Google, JULES is not the only artificial intelligence coding platform for building applications. Google provides symbol, AI Studio, JULES and Firebase.

Firebase, which was announced in April, allows non -visible applications to build and add artificial intelligence features. Google has updated the arterial system, with the addition of a new work space of artificial intelligence to Firebase Studio and Firebase Ai Logic to monitor the use of artificial intelligence.

Firebase Studio, backed by Gemini 2.5 Pro, allows users to build more advanced applications. Firebase Ai Logic provides developers a way to add features to the back interface of the application, such as ratification and identity. People are also allowed to verify the use of the distinctive symbol or solve cumin problems without the need for a third -party coincidence program.

Janine Banks, Vice president and Director General of the developer X and head of developer relations in Google, Venturebeat that Firebase distinguishes himself from Google’s jols and other Google coding products by being the first place for new people in coding the experience of making their AI applications.

“Google provides many great tools to help you in specialized parts of your affirmation. So, for example, you can use Google Ai Studio, which helps in trying to infer artificial intelligence to know the best improved claims.” “But Firebase is the only place to merge all these things together, and it is one place for the developers and professionals of full gains, but also creators who get coding.”