Grok Build Under Fire: Developers Compare It to Claude Code and Codex and Are Left Disappointed
Elon Musk, known for his love of bold statements, recently reached out to the developer community with a direct request: "Tell me what's wrong with Grok Build." The response was, to say the least, unexpected. Instead of enthusiastic reviews, he received a barrage of constructive, and at times harsh, criticism. Let's break down exactly what users dislike about the new CLI tool from xAI and why it currently lags behind giants like Claude Code and OpenAI Codex.
What is Grok Build?
Grok Build is an agentic development tool launched in beta in May 2026. It runs from the terminal and is exclusively available to SuperGrok and X Premium Plus subscribers. Access costs around $300 per month, automatically placing it in the same price range as Claude Code and GitHub Copilot. For this price, users get access to the Grok 4.3 beta model with a 16-agent architecture and a 2 million token context window. The tool also supports a planning mode where each step can be approved or modified before execution, and all changes are displayed as a diff.
Comparison with Competitors: Not in xAI's Favor
The main theme running through all the comments was comparison with direct competitors. And here, the metrics are discouraging. One developer shared a personal experiment: Grok Build spent nearly two days on a task that Codex solved in six hours, while advancing twice as far. Another user complained about endless 30-minute cycles, while Opus from Anthropic solved the same problem on the first try. A third stated that the inference speed in Grok CLI is uncomfortably slow compared to Claude Code and Codex.
Ultimately, developers concluded: Grok Build might be good for deep research, but in complex autonomous coding, it clearly falls short of its rivals.
Feature and Ecosystem Requests
Beyond code quality, the community actively pointed out ecosystem shortcomings. The main request was the creation of a desktop application similar to Claude Cowork. Developers note that Claude's strength lies in its integration into all workflows, not just code writing. Other requests include: releasing an open-source version during the beta testing phase, implementing full loop skill functionality, creating a /goal command for stable autonomous operation, and integrating a built-in demo of the software being built. The issue of feedback was also raised separately: one user admitted they didn't know through which channels to send feedback, even though the /feedback command is already built into the CLI.
Price and Limitations: Expensive and Strict
The high subscription cost sparked a separate wave of criticism. Developers complain about the rigid tie to the expensive SuperGrok plan and suggest introducing a more affordable tier. Additionally, they lament strict token limits and only 15 minutes of Grok Premium access per day. There are also concerns that xAI might repeat Claude's withdrawal from Europe, exacerbating geographic restrictions.
Conclusions: The Gap Between Marketing and Reality
The collection of comments under Musk's post revealed 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, it loses in quality, speed, and stability. The key complaints boil down to three areas: quality (cycles, regressions, losing in direct tests), ecosystem (lack of a desktop app, open-source, and several agentic features), and economics (high price and strict limits).
My analytical assessment: The early beta stage and built-in feedback mechanism give xAI a direct channel for rapid product refinement. However, to catch up with the leaders, the company will need not just to fix bugs, but to fundamentally rethink its pricing strategy and ecosystem development. Otherwise, Grok Build risks remaining a niche tool for enthusiasts rather than a serious competitor in the AI coding market.