Do Claude Code, Codex and Cursor pick the same tools?

If you measure how coding agents choose your product and you only measure one agent, you know about half of your market.

We ran the same requests through Claude Code, Codex and Cursor, inside the same repositories, 11,600 times, as part of 15,287 judged sessions across five agents. Here is how often those three agreed.

The headline

The full matrix

What agreement looks like

These seven categories have leaders with at least 41% of sessions, and all three agents agree on their leaders.

When a product holds this much of a category, every agent reaches for it without much deliberation. It is in the training data, in the templates, and in the repositories.

Agreement is not a sign that the agents are thinking alike. It is a sign that there is nothing left to decide.

The other five agreements are thinner. In internationalization all three put next-intl first, but it holds 22% of sessions.

What disagreement looks like

The categories where they differ are the ones with no runaway leader. In search, the top product holds 16%. In agent frameworks, 14%. In AI gateways, 26%.

In those categories the agent is genuinely choosing, and small differences in how each agent works produce different answers.

If you are deciding where to spend as a result of this, which coding agent should you optimise for? works through it.

Why they disagree

Three general causes. A fourth, the agent's own maker, has its own section below.

They search the web at very different rates

This is the biggest single difference between Claude Code and Codex, and it is large. Codex reads the current web in almost every session. In seven sessions out of ten, Claude Code works from what it already knows plus what is in the repository.

The practical consequence: a page you publish can reach Codex in almost every session and Claude Code in about three in ten. If your users run Claude Code, content is a weaker lever and the repository a stronger one.

They differ on building it themselves

Claude Code reaches for a hand-written implementation more than twice as often as Codex, and more often than Cursor. If you sell in a category where hand-written code is a plausible answer, Claude Code is a harder market, and the argument you have to win is not "why us over them" but "why a product at all".

Their training data and defaults differ

Every model carries a different set of habits. In email, Claude Code put Postmark first while both others put Resend first. In sandboxes, Codex put Modal first in 91 of its 170 sessions while both others put E2B first. Each of those rests on more than 80 sessions per agent, not a handful.

The self-preference question

In code review, Claude Code and Cursor each put its own maker's product first: Claude Code chose Claude Code review in 101 of 130 sessions and Cursor chose Cursor Bugbot in 91 of 120. Codex chose Codex code review in 44 of 160 sessions, second to CodeRabbit with 46. Claude Code and Cursor never chose a rival maker's reviewer. Codex chose Claude Code review once.

It shows in other categories too. Cursor put the Cursor SDK first in agent frameworks and Cursor Automations first in AI SRE. Claude Code put Anthropic web search first in AI search and Anthropic Claude first in document processing. That is five of the 17 categories where the three disagree.

It is worth being careful here. These are real products in their categories, and a reasonable choice inside the maker's own environment. The prompts in those categories do not mention any agent maker, except two agent framework prompts that name Claude and ChatGPT together, in alternating order, with a linter that checks it.

So the honest statement is: where an agent's maker sells a product in the category, the agent often puts it first. It is not universal. Codex chose OpenAI models in 7 of 235 document processing sessions. See self-preference for every case we measured.

What this means if you sell software

Measure per agent, always. A blended install share is an average across markets that behave differently. It will tell you that you hold 22% of a category while hiding that you hold 40% with one agent and 5% with another. The fix for those two situations is not the same.

Match the lever to the agent. Content and search visibility move Codex. The repository, the templates and the training data move Claude Code. Find out which agent your users run before choosing where to spend.

Treat a split category as an opportunity. If the three agents disagree, the category is not settled. Nothing is locked in and the ordering can be moved. The exception is a split where agents pick their own maker's product, as Claude Code and Cursor do in code review. If all three agree and the leader holds 60%, you are not going to change that with a blog post this quarter, and you should be looking for the requests where the leader fits badly.

Watch the in-house rate per agent. Losing to Claude Code writing its own implementation is a completely different problem from losing to a competitor, and it needs different content: an argument that the problem is harder than it looks, with specifics.

How this was measured

15,287 judged sessions across 31 categories and five agents, 11,600 of them from Claude Code, Codex and Cursor. Real command line tools at pinned versions, in sandboxes, inside 110 realistic codebases with real lock files and replayed git histories. Requests written as four kinds of person, in several wordings. A simulated project owner approves or questions each recommendation before any code is written. A judge from a model family that builds none of the agents reads every session blind.

Three copies of the judge agreed with each other 97.9% of the time.

Every run is published and can be replayed on the sector boards.