Grok Build Under Fire: xAI Loses the Battle to Claude Code and Codex
Elon Musk publicly requested feedback on the new xAI tool — Grok Build. The community's reaction was not just mixed, but openly critical. Dozens of developers directly compared it to competitors, and the results for xAI look discouraging. Let's break down what exactly went wrong.
What is Grok Build and what is its problem
Grok Build is an agentic CLI development tool launched in May 2026. It is available only to SuperGrok and X Premium Plus subscribers for approximately $300 per month. The base model Grok 4.3 beta uses an architecture with 16 agents and a context window of 2 million tokens. It sounds impressive, but in practice, things turned out differently.
Developers conducted direct experiments. One showed that Grok worked on implementing a project for nearly two days, while OpenAI Codex completed the same volume of tasks in six hours, advancing twice as far. Another user reported that Grok went into infinite loops for thirty minutes, while Opus from Anthropic solved the problem on the first try. Inference speed also leaves much to be desired — watching the agent work is simply uncomfortable.
Functional gaps and price
Users are actively requesting the creation of an official desktop application similar to Claude Cowork. They note that Claude's main strength lies in its integration into all aspects of workflows, not just code writing. Additionally, there are calls for an open-source version, the implementation of full loop functionality, and the integration of a built-in capability to demonstrate the software being created.
The $300 per month price tag sparked a separate wave of criticism. Developers complain about the strict tie to the expensive SuperGrok plan and suggest introducing a more affordable tier. Also noted are strict token limits and a daily cap of 15 minutes of access to Grok Premium. Concerns are raised that xAI might repeat Claude's withdrawal from Europe.
Irony and skepticism
The very format of Musk's post sparked sarcastic reactions. The request for critical feedback was accompanied by a quote from an enthusiastic fan who literally professed love for the product. One commenter called reposting one's own praise a special kind of self-confidence.
At the same time, part of the audience remained loyal. There were thanks for the team's rapid iterations and statements that the product is improving quickly. Some even predict that Grok will soon become the best tool on the market.
Analyst conclusions
The collection of reviews demonstrates an obvious gap between the marketing message and the assessments of practicing developers. Grok Build stands out with its large context and multi-agent architecture, but in real-world tasks, users note a lag in the quality of autonomous coding, speed, and stability compared to established Claude Code and Codex. The early beta stage and built-in feedback mechanism give xAI a direct channel for rapid refinement, but at this point, the product clearly does not meet the claimed level.
My expert opinion: xAI is making a classic mistake by assuming that raw model power can compensate for a lack of polished user experience and ecosystem. In the AI agent race, the winner is not the one with more tokens, but the one offering a reliable, fast, and intuitive tool. For now, Grok Build is a promising but immature prototype that has a long way to go before becoming a real competitor to market leaders.