Grok Build vs Claude Code vs Codex in 2026: which coding agent to use, and when

Disclosure: Enjoy is our product, and it runs all three agents. We checked every fact below against SpaceXSI’s, Anthropic’s, OpenAI’s and Artificial Analysis’s own pages on October 30, 2026, and we say where you don’t need Enjoy at all.

Grok Build vs Claude Code vs Codex, compared

All three do the same job: you describe a change, and the agent reads your project, edits files and runs commands until it’s done. Grok Build comes from SpaceXSI, the company behind Grok; Claude Code from Anthropic; Codex from OpenAI. Each runs on its own company’s plans, so choosing one mostly means choosing which subscription to pay for.

Option	Best for	Price, checked October 30, 2026	Runs on	Open source
Claude Code	The hardest work	Claude Pro $20 a month, or $17 billed yearly; Max $100; Ultra, formerly Max 20x, $200	Mac, Windows, Linux	No
Codex	The lowest cost per task, and a free start	Free and Go ($8) in the desktop app, with GPT-6 Luna; Plus $20; Pro $100, $200 or $500	Mac, Windows, Linux	The CLI, Apache 2.0
Grok Build	People who already pay for Grok or X, from $8	Free, limited; Plus $8; Premium $30; Super $100; Ultra $200	Mac, Windows, Linux	Yes, Apache 2.0
Enjoy	All three in one conversation	A monthly subscription on top of your plans, free for 7 days (pricing)	Mac, Windows, Linux	No

Prices are each company’s US list prices per month, before tax. All three agents can also run on an API key that bills per token instead of a plan.

How the three score

	Claude Code	Codex	Grok Build
Coding Agent Index, best setup	62, with Opus 5.5 at max effort	55, with GPT-6 Astra at max effort	47, with Grok 4.7 at extra high effort
API cost per task	$11.42	$7.04 with GPT-6 Astra; $1.50 with GPT-6.1 Sol, which scores 54	$9.48
Time per task	1.0 hour	24.8 minutes with GPT-6 Astra; 23.7 with GPT-6.1 Sol	35.8 minutes
Tokens per task	14.5 million	3.2 million with GPT-6 Astra	15.9 million

The figures come from Artificial Analysis’s Coding Agent Index, which runs each agent and model together on 290 tasks from FrontierCode, Terminal-Bench and SWE-Atlas, as its Claude Code vs Grok Build and Codex vs Grok Build pages stood on October 30. Grok Build’s gap is widest on Terminal-Bench, at 33% against Claude Code’s 63% and Codex’s 56%; on SWE-Atlas’s technical questions it scores 63%, ahead of Codex’s 59% and close to Claude Code’s 66%. The costs are “pay-per-token API cost, not consumer plan pricing”: on a plan, a task draws on your limits instead.

Claude Code

Best for: the hardest work, when results matter more than cost.

Claude Code runs in a terminal on macOS, Windows and Linux, and in Anthropic’s desktop app, editors and browser. It has the top score on the index, and when a limit stops it partway through a task, its docs say “Claude Code waits in the open session and continues the task on its own after the limit resets.”

Price: “Claude Code is included in all paid plans,” says Claude’s pricing, and the Free plan leaves it out. Pro is $20 a month, or $17 billed yearly; Max is $100, and Ultra, called Max 20x until October 30, is $200. Fable 5.1 comes with Max and Ultra, up to “50% of weekly limits,” and bills to usage credits on Pro.

Watch for: the limits are shared with your Claude chats, so “your work in the terminal and your chats draw from one pool,” and Anthropic publishes no message counts: “there’s no fixed message count.” It’s the slowest of the three on the index, at an hour a task. Claude Code isn’t open source: its license reads “© Anthropic PBC. All rights reserved.” And on native Windows, its setup guide lists sandboxing as “Not supported.”

Codex

Best for: the lowest cost per task, a free start, and Windows.

Codex runs in the ChatGPT desktop app, in a terminal, in your editor and in OpenAI’s cloud. The desktop app is for “macOS, Windows, or Linux,” and on Windows it “runs natively on Windows using PowerShell and the Windows sandbox,” per its Windows page. The Codex CLI is open source under Apache 2.0. On the index, GPT-6.1 Sol scores 54 for $1.50 a task, less than a sixth of Grok Build’s cost for 7 points more.

Price: Free, and Go at $8 a month, get “GPT-6 Luna at Standard speed in the desktop app, subject to rollout,” per Codex’s pricing. Plus, at $20, adds GPT-6.1 Sol, which OpenAI estimates at 15 to 160 local messages every five hours, and Pro is $100, $200 or $500 a month, where “Pro plans currently have no five-hour limit.”

Watch for: OpenAI says its counts “are not fixed message limits,” and Codex with GPT-6 Luna, the free plans’ model, scores 31 on the index, far below Sol. The plans move, too: on October 29 OpenAI canceled a cut to Pro 200 it had announced a month earlier.

Grok Build

Best for: people who already pay for Grok or X, and anyone who wants an open-source agent with a free start.

Grok Build is “a powerful and extensible coding agent,” its docs say, used “via an interactive TUI, headlessly in scripts or bots, or through the Agent Client Protocol (ACP) in other apps.” It installs with one command on Mac and Linux, or a PowerShell script on Windows, and is now “powered by Grok 4.7.” It has a plan mode, subagents that can each run in their own worktree, skills, plugins, hooks and MCP servers, and it reads Claude Code’s files: “AGENTS.md, Agents.md, AGENT.md, CLAUDE.md, Claude.md, and CLAUDE.local.md,” per its project rules page. “Grok supports any custom model,” with your own API key for it. Its code is open under Apache 2.0, and its changelog lists version 1.0.70, released October 23.

Price: Grok’s pricing is one subscription for Grok, X and Cursor. Free lists Grok Build as “Limited.” Plus, at $8 a month or $80 a year, includes “Grok Chat, Voice, Imagine, and Build,” the Grok API and Cursor. Premium, at $30, has “4x more usage compared to Plus”; Super, at $100, 15 times; and Ultra, at $200, 50 times. Grok 4.7 costs $2 per million input tokens and $6 per million output on the API, per its announcement.

Watch for: Plus’s usage is a “Shared AI usage pool” across Grok’s products, so chats, images and Cursor draw on the same allowance as Grok Build. Premium lists “Latest Grok models” among what it adds to Plus, so check which models Plus gets. On the index it trails both rivals, and its tasks used 15.9 million tokens and 170 turns on average, against Codex’s 3.2 million and 39. And its sandbox is “Off by default”; the docs name Landlock on Linux and Seatbelt on macOS, and none for Windows.

Enjoy

Best for: using all three in one conversation, and keeping going when one runs out.

Enjoy runs Grok Build, Claude Code and Codex, along with Antigravity, Muse Code and local models through OpenCode, on Mac, Windows and Linux, each signed in with the plan you already have. What it adds for these three:

• One conversation, any agent. Switch agents at any point: “The new agent gets the conversation so far, including replies from other agents, before it works on your message.” Plan with Claude Code, build with Codex, and have Grok Build review it.
• Limits handled. When an agent runs out, Enjoy switches to another of your accounts with usage left, or waits for the reset and carries on. Claude Code, Codex and Grok Build can each keep several accounts signed in.
• Usage in view. It shows what each signed-in account has left, with one bar for Grok Build, named Weekly or Subscription.

Price: Enjoy is one monthly subscription, and the first seven days are free, with a card to start. It includes your phone and any browser and your other computers, and each teammate you invite needs a subscription of their own. See pricing.

Watch for: Enjoy doesn’t add one lab’s models to another lab’s plan: each agent runs on its own company’s subscription. Grok Build doesn’t tell Enjoy which account is which, so its accounts are numbered, and it has no Fast mode. Waiting for a reset needs “Your computer must be on, with Enjoy open, when the usage resets,” and scheduled workflows also run only while it’s awake with Enjoy open. Enjoy isn’t open source, and it’s a monthly subscription on top of your AI plans once the free week ends.

When one agent is enough

If you already pay for one and rarely run out, a second plan buys little. Stay with Claude Code when its results are why you pay, and wait out a session limit or buy usage credits when you hit one. Stay with Codex when cost or Windows matter most. Stay with Grok Build when your Grok or X subscription already covers it and the work isn’t the hardest kind: at $8 a month with Cursor included, it’s the cheapest paid start of the three.

Add a second agent when one plan’s limits stop you most weeks, or when you want another lab’s review of the same code. Our Claude Code vs Codex guide covers that pairing in more depth.

How we compared

We read SpaceXSI’s Grok Build page, docs, changelog, installer, GitHub repository, pricing, plans and Grok 4.7 announcement; Anthropic’s pricing, Claude Code docs and license; OpenAI’s Codex and ChatGPT pricing and desktop app docs; and Artificial Analysis’s index and comparison pages, all on October 30, 2026, and every one is linked below. We quote each company’s own words for its limits and leave out what it hasn’t documented. Artificial Analysis rounds scores on its comparison pages, so the scores here are whole numbers. Plans and models change often, so we recheck this guide against its sources at least once a month, and the date at the foot of the page says when we last did.