No single best tool. Three that differ by where the agent runs.
There is no single best AI coding tool; there are three that differ by where the agent runs. Cursor keeps it in a VS Code-based editor. Claude Code starts in the terminal and documents 33 hook events. Codex spreads one agent across the CLI, IDE extension, desktop app, cloud tasks and Slack and Linear hand-offs. All three share most capabilities by name. AI Developer Toolkit teaches all three tools side by side: 400+ guides and 100+ recipes, each showing when it was last updated.
$59.99 once for 12 months of access. Nothing renews on its own. 30-day refund on a first purchase.
- Compare data, updated
- 2026-10-16
- Cursor docs
- 2026-08-28
- Claude Code 2.1.283
- 2026-09-26
- Codex 0.157.1
- 2026-09-26
- List prices captured
- 2026-09-26
Which one fits how you work?
Tick the lines that are true for you. Each one is a verdict line from the toolkit’s compare data, checked 2026‑10‑16.
Pick Cursor if
Editor-first: you steer each change as it lands
Pick Claude Code if
Terminal-first: you review the result
Pick Codex if
Multi-surface: one agent across CLI, IDE, desktop and cloud
Station by station.
The jobs a coding-agent workflow has to do, with what each tool names for them. The note beside each station says whether the difference should change your choice.
Intent
Project instructions, skills, plan mode
Same shape All three read a project instruction file, install skills and plan before code; the file name changes: CLAUDE.md in Claude Code, AGENTS.md in Codex, Rules and AGENTS.md in Cursor.
Cursor
- Rules
- Cursor's docs:
AGENTS.mdin the project root and in subdirectories (nested files combine, the more specific one wins), andCLAUDE.mdread the same way and always applied;.cursorrulesis legacy Cursor token management guide, 2026-10-12 - The agent runtime in
@cursor/sdk1.0.32 loads a root and nestedAGENTS.mdunconditionally, andCLAUDE.mdonly with third-party config loading on, while Cursor's docs sayCLAUDE.mdis always applied Cursor token management guide, Cursor SDK 1.0.32, 2026-09-26 - Agent Skills
- Plan Mode
Cursor docs, 2026-08-28
Claude Code
-
CLAUDE.md; by default readsAGENTS.mdonly when there is noCLAUDE.md(v2.1.277+), and the/configProject instructions setting loads both - Skills in
.claude/skills/; bundled/code-review,/batch,/simplify - Plan mode:
/plan, Shift+Tab,--permission-mode plan
Claude Code 2.1.283, 2026-09-26
Codex
-
AGENTS.md(/initcreates one) - Skills:
/skills,.agents/skills/ - Plan:
/plan
Codex 0.157.1, 2026-09-26
Harness
Permissions and approvals, sandbox, MCP, hooks, organisation policy
Differs Enforcement depth differs: 33 documented hook events in Claude Code against 12 in Codex, and Read deny rules in Claude Code's settings; Cursor's hooks can block, its agent runtime refuses reads of .cursorignore paths, and its run modes are Auto-review, Allowlist and Run Everything (Cursor docs, checked 2026-10-12).
Cursor
- Run modes: Auto-review (the recommended default), Allowlist and Run Everything Cursor docs, Cursor 3.24, 2026-10-12
- Auto-review runs shell commands in a sandbox when possible;
sandbox.jsonconfigures it Cursor docs, Cursor 3.24, 2026-10-12 - Hooks: JSON over stdio, before or after agent-loop stages; can block
- MCP
- Organisation policy: Team and MDM settings layers (SDK 1.0.32)
- File deny rules: the agent runtime refuses reads of paths in
.cursorignore, passes the list to the shell sandbox and blocks the agent from editing the file Cursor privacy and security guide, Cursor SDK 1.0.32, 2026-09-22 - Starting mode: Auto-review, the recommended default since 3.6 (2026-05-29) Cursor docs, Cursor 3.24, 2026-10-12
Cursor docs, 2026-08-28
Claude Code
- Permission modes: Manual (
default),acceptEdits,plan,auto,dontAsk,bypassPermissions; interactive sessions start in auto mode from v2.1.283,claude -pin Manual (auto mode there only on third-party providers or with telemetry off, from v2.1.285) - Sandboxed Bash: macOS Seatbelt, Linux and WSL2 bubblewrap
- Hooks: 33 events;
command,http,mcp_tool,promptandagenthandlers - MCP client:
claude mcp add(--scope projectwrites.mcp.json); runs as an MCP server withclaude mcp serve -
Readdeny rules in.claude/settings.json, for exampleRead(./.env) - Organisation policy: managed settings (
availableModels,permissions.disableAutoMode,deniedModels; v2.1.283)
Claude Code 2.1.283, 2026-09-26
Codex
- Approval policy (
-a on-requestor-a never) plus automatic review (--approve-for-me); starts inon-request - OS sandbox:
--sandbox read-only,workspace-write,danger-full-access; permission profiles (beta) - Hooks: 12 events; need persisted trust before they run
- MCP client:
codex mcp add,config.toml; does not run as an MCP server (removed in 0.154.0) - File access: a permission profile with a
denyfilesystem rule (beta);--sandboxalone restricts writes, not reads - Organisation policy:
requirements.toml(allow_managed_hooks_only,permission_profile,mcp_servers)
Codex 0.157.1, 2026-09-26
Loop
Stop conditions, /goal, /loop, scheduled runs, headless
Differs All three run headless and on a schedule, but Claude Code Routines and Cursor Automations run in the cloud while Codex automations run locally in the desktop app. /goal and /loop differ (Codex has no /loop in 0.162.1; Cursor's /goal is marked "Rolling out"), and Claude Code can also serve as an MCP server, which Codex removed in 0.154.0.
Cursor
-
/goal, marked "Rolling out" Cursor docs, Cursor 3.24, 2026-10-12 -
/loopbundled skill since 3.5 - Scheduled runs: Automations, which start Cloud Agents
- Headless: print mode (
agent -p) Cursor docs, Cursor 3.24, 2026-10-12 - Machine-readable output:
--output-format textis documented Cursor docs, Cursor 3.24, 2026-10-12 - Other output formats: not verified for this page (checked 2026-10-12) Cursor docs, Cursor 3.24, 2026-10-12
Cursor docs, 2026-08-28
Claude Code
-
/goal, judged by a separate evaluator model each turn -
/loopbundled skill (alias/proactive) - Scheduled runs: Routines in the cloud (
/schedule, research preview) - Headless:
claude -p;--output-format jsonorstream-json;--json-schema
Claude Code 2.1.283, 2026-09-26
Codex
-
/goal(featuregoals, stable) - No
/loopin Codex 0.162.1 (slash-command enum, binary strings) or the developer-commands reference (checked 2026-10-12); use/goalor CI Loop command guide, Codex 0.162.1, 2026-10-12 - Scheduled runs: automations run locally in the desktop app, which must be open and the machine awake Background and cloud agents guide, 2026-09-26
- Headless:
codex exec(aliase);--json(JSONL events);--output-schema FILE;-o FILE
Codex 0.157.1, 2026-09-26
Graph
Subagents, worktrees, background sessions
Differs Where the agent runs differs: in the editor (Cursor), from the terminal out to IDE, desktop and web (Claude Code), or spread across CLI, IDE, desktop app, cloud tasks and Slack and Linear hand-offs (Codex).
Cursor
- Subagents, including on their own cloud VMs
- Worktrees
- Background sessions: Cloud Agents in isolated VMs
Cursor docs, 2026-08-28
Claude Code
- Subagents: built-in Explore, Plan and general-purpose; custom ones in
.claude/agents/ - Worktrees:
--worktree [name](-w) - Background sessions:
--bg, thenclaude attach,logs,stop - Cloud: web sessions (
claude --cloud)
Claude Code 2.1.283, 2026-09-26
Codex
- Subagents:
/subagents; multi-agent on by default - Worktrees:
--worktreeflag or/worktree - Background sessions: the
codex agentscommand center - Cloud: Codex cloud (
codex cloud, experimental)
Codex 0.157.1, 2026-09-26
Verification
Diff review, review agents, security review, checkpoints and rewind
Same shape All three review in the session and rewind. Claude Code has /code-review, /security-review and checkpoints; Codex has /review; Cursor has Agent Review, /review and /review-security. One caveat: Codex's Esc Esc backtracks the conversation but does not restore files (0.157.1), so file state is Git's job.
Cursor
- In the session: Agent Review (
/agent-review) and/review, which runs/review-bugbotCursor code review guide, 2026-09-15 - Bugbot on the pull request
- Security review:
/review-securitybefore the push; Security Agents on the pull request Cursor code review guide, 2026-09-15 - Rewind: checkpoints, stored locally and separate from Git; restoring one reverts files only Cursor docs, Cursor 3.24, 2026-10-12
Cursor docs, 2026-08-28
Claude Code
- In the session:
/code-review(--fix,--comment) - From the shell or CI:
claude ultrareview(cloud, multi-agent); managed Code Review (Team and Enterprise; generally available per Anthropic) Code review automation guide, 2026-10-16 - Security review:
/security-review - Rewind: checkpoints and
/rewind(Esc Esc)
Claude Code 2.1.283, 2026-09-26
Codex
- In the session:
/review - From the shell or CI:
codex review,codex exec review - Security review:
@codex security review - Rewind: Esc Esc backtracks the conversation to an earlier message; it does not restore files (0.157.1). Use Git for file state
Codex 0.157.1, 2026-09-26
Release
CI and GitHub Actions, repo-event triggers, pull-request routing
Differs Review on the pull request differs: Codex reviews on a mention (@codex review), Cursor adds Bugbot, Security Agents and reviewer routing, Claude Code reviews through its GitHub Action or managed Code Review.
Cursor
- GitHub Actions: Cursor CLI in GitHub Actions (documented)
- Review on the pull request: Bugbot
- Hand-offs: Automations triggered from Slack, Linear, GitHub, GitLab, webhooks, Sentry and PagerDuty
- Reviewer routing and approval: PR Routing & Approval (can approve low-risk PRs)
- Start a cloud run from code: Cloud Agents API and webhooks
Cursor docs, 2026-08-28
Claude Code
- GitHub Actions and review on the pull request:
anthropics/claude-code-action@v1 - Hand-offs: Slack (Claude Tag for Team and Enterprise); others through MCP
- Reviewer routing and approval: through your own CI and CODEOWNERS
- Start a cloud run from code: Routines with an API trigger (research preview)
Claude Code 2.1.283, 2026-09-26
Codex
- GitHub Actions:
openai/codex-action@v1 - Review on the pull request:
@codex reviewon GitHub and GitLab, on a mention and without a separate workflow file - Hand-offs: Slack (
@ChatGPT, with Cloud delegation to a Codex Cloud task) and Linear (@Codex) Codex docs, 2026-10-12 - Reviewer routing and approval: through your own CI and CODEOWNERS
- Start a cloud run from code:
codex cloud(experimental)
Codex 0.157.1, 2026-09-26
Editor
Tab completions, inline edit, extensions and keybindings
Differs Tab completion and inline edit are documented for Cursor. The docs describe Claude Code and Codex as agents, not completion engines.
Cursor
- The Cursor editor itself, with Tab completion and inline edit
- Inline edit is usually bound to
Cmd+K(macOS) orCtrl+K(Windows, Linux); shortcuts move between builds Inline edit guide, 2026-09-26 - Shortcuts are VS Code keybindings attached to Cursor's AI commands; pin them in
keybindings.jsonCursor shortcuts reference, 2026-09-26 - VS Code extensions IDE migration guide, 2026-09-26
Cursor docs, 2026-08-28
Claude Code
- VS Code and JetBrains extensions: agent panel, diffs, checkpoints
- Tab completions and
Cmd+Kinline edit: not found in the docs (checked 2026-09-26). The migration guide calls Claude Code an agent, not a completion engine, and keeps your editor's completion beside it IDE migration guide, 2026-09-26
Claude Code 2.1.283, 2026-09-26
Codex
- IDE extension for VS Code, Cursor and Windsurf (per the OpenAI README)
- Tab completions and inline edit: not found in the docs (checked 2026-09-26). The migration guide calls Codex an agent, not a completion engine, and keeps your editor's completion beside it IDE migration guide, 2026-09-26
Codex 0.157.1, 2026-09-26
Read a gap as a gap in what we read. A line that says “not found in the docs” means we could not source it on that date. A Cursor cell with no later stamp comes from docs read before cursor.com could not be reached on 2026‑09‑26. Re-check a Cursor cell before you automate against it.
- Drawing
- Cursor vs Claude Code vs Codex
- Source
- The compare data, 2026‑10‑16. Each cell above names the page and the day it was read.
Models and what they cost.
Claude Code is included in Claude Pro at $20/month, Max 5x at $100 and Max 20x at $200. Codex is included in ChatGPT Plus ($20/month), Pro ($100, $200 or $500) and Business ($20 per user per month billed annually), and in Edu and Enterprise (OpenAI, read 2026-10-12); the 5x and 20x Plus-usage ratios of Pro at $100 and $200 come from a secondary source (TechCrunch, 2026-04-09). Cursor Pro is $20/month, Pro+ $60 and Ultra $200 (cursor.com/pricing, read 2026-10-12).
| US list prices, 2026‑09‑26 | Cursor | Claude Code | Codex |
|---|---|---|---|
| Models | Several vendors in one picker. No single default model: Hobby uses Auto. Cursor's own pool is Grok 4.7 (which Cursor calls its flagship model), Grok 4.6, Grok 4.5 and Composer 2.5; Claude Opus 5.5, Fable 5.1, Sonnet 5.5 and Haiku 5.5 are third-party models billed at API rates; no GPT-6 model is listed (Cursor, checked 2026-10-12) | Default Claude Opus 5.5 (v2.1.280+, on both release channels on 2026-10-17: stable v2.1.292, latest v2.1.302). Claude Fable 5.1 by hand with /model fable; Claude Sonnet 5.5 and Claude Haiku 5.5 when cost dominates (the haiku alias needs v2.1.293, latest channel only on 2026-10-17). Switch mid-session with /model | Default GPT-6.1 Sol (since CLI 0.159.1, 2026-09-29, at low effort); GPT-6 Astra chosen by hand for the hardest work; GPT-6 Sol and GPT-6 Luna in the picker; GPT-5.6 Sol, Terra and Luna still selectable. Switch with /model |
| Context window | Set by the model: Grok 4.7 has a 256K standard window and 500K long context, and Claude Sonnet 5.5 up to 1M (Cursor model pages, checked 2026-10-12) | 1M tokens on Claude Opus 5.5, Fable 5.1, Sonnet 5.5 and Haiku 5.5 | 272K by default in Codex, up to 872K (GPT-6.1 Sol, GPT-6 Astra, Sol and Luna) |
| Entry plan | $20/month Cursor ProUnlimited Tab completions, extended Agent usage limits on all models, Bugbot and Cloud Agents (cursor.com/pricing, read 2026-10-12) | $20/month Claude ProOr $17/month billed annually ($200 up front); Claude Code included | $20/month ChatGPT PlusCodex included (OpenAI pricing page, read 2026-10-12) |
| Middle plan | $60/month Cursor Pro+The same individual-plan features as Pro; Cursor recommends Pro+ for daily Agent users (cursor.com/pricing, read 2026-10-12) | $100/month Claude Max 5xClaude Code; five times Pro usage | $100/month ChatGPT Pro (5x Plus)Codex; 5x Plus usage on Codex (secondary source: TechCrunch, 2026-04-09). Price: OpenAI, read 2026-10-12 |
| Top individual plan | $200/month Cursor UltraThe same individual-plan features as Pro; Cursor recommends Ultra for Agent power users (cursor.com/pricing, read 2026-10-12) | $200/month Claude Max 20xClaude Code; twenty times Pro usage | $200/month ChatGPT Pro (20x Plus)Codex; 20x Plus usage on Codex (secondary source: TechCrunch, 2026-04-09). Price: OpenAI, read 2026-10-12. OpenAI also lists a $500 Pro tier |
| Team seat | Teams Standard $40/user/month ($32 billed annually); Teams Premium $120/user/month ($96 billed annually), "5x the usage of a Standard seat"; Enterprise custom (cursor.com/pricing, read 2026-10-12) | Team Standard $25/seat monthly ($20 annual); Team Premium $125/seat monthly ($100 annual); Enterprise $20/seat/month billed annually, plus usage at API rates | ChatGPT Business $20/user/month billed annually ($25 billed monthly, 2+ users), with Codex included; Edu and Enterprise include Codex, price on contact (OpenAI, read 2026-10-12) |
Use both, if each one earns its seat
An editor-based tool and a CLI agent take different parts of a day: one for the change you are typing, the other for the change you describe and come back to.
The cost is real: every seat is a separate subscription, so ask which capability the second tool has that the first lacks before you add it.
Hold them to one gate: tests, type check and lint in CI on every pull request, whichever agent opened it, with one review bot per repository.
$59.99 once for 12 months of access. Nothing renews on its own. 30-day refund on a first purchase.
In the editor
A small edit in the file you have open
Tab and inline edit keep you in the file; there is nothing to delegate.
On screen
A UI change you judge by looking at it
You accept or reject each change as the dev server reloads.
Across the repo
A change across 10 or more files, with tests as the oracle
It runs long in a worktree, and its richer hooks enforce the gate after every edit.
In scripts
A scripted job or a CI step
claude -p --output-format json returns a result CI can read.
From chat
A bug reported in a Slack thread or a Linear issue
@ChatGPT in Slack (with Cloud delegation) or @Codex in Linear starts a Codex cloud task from the thread; nobody changes tools.
On the pull request
Review of every pull request
@codex review with automatic reviews, or Bugbot; run one bot per repository, not both.
CLAUDE.md, AGENTS.md, Rules); keep them in sync, install skills with npx skills add <owner/repo>, and hold every tool to the same gate in CI. “Using several of them can feel messy. I switch between Claude Code, Cursor, VS Code, GitHub, and other agents, and often have to explain the same project or problem again.” In more detail.
Which AI coding tool should I start with?
Start from the subscription you already have. Claude Code is included in Claude Pro, Max, Team and Enterprise plans; Codex is included in ChatGPT Plus, Pro, Business, Edu and Enterprise. If you pay for neither, start from where you want to work: an editor points to Cursor, a terminal points to Claude Code, and GitHub, Slack or Linear points to Codex.
Questions before you choose.
Something else? Write to support@developertoolkit.ai.
Which AI coding tool is best in 2026?
None of the three is best at everything. On 2026-09-26 they share subagents, worktrees, skills, hooks, MCP, a headless mode, an SDK and scheduled runs (cloud Routines in Claude Code, cloud Automations in Cursor, local desktop-app automations in Codex), and differ in where the agent runs, how guardrails are enforced and what reviews a pull request. Test the rows that decide your choice on your own repository.
Do all three support MCP servers?
Yes. Claude Code uses claude mcp add or .mcp.json, Codex uses codex mcp add and config.toml, and Cursor has MCP. All three also install Agent Skills the same way, with npx skills add <owner/repo>.
How much do they cost?
Claude Code is included in Claude Pro at $20/month, Max 5x at $100 and Max 20x at $200; Team Standard is $25 a seat and Team Premium $125 ($20 and $100 billed annually). Codex is included in ChatGPT plans: Plus is $20/month, Pro $100, $200 or $500, and Business $20 per user per month billed annually (OpenAI, read 2026-10-12); the 5x and 20x usage ratios of Pro at $100 and $200 come from a secondary source. Cursor Pro is $20/month, Pro+ $60 and Ultra $200, and Teams Standard and Premium are $40 and $120 per user per month (cursor.com/pricing, read 2026-10-12). Claude prices are US list prices read on 2026-09-26.
Can one tool replace a developer?
No. All three plan, edit files and run commands, and a change still needs a gate: the docs keep the merge decision on your own tests and CI, not on the agent or a review bot. Treat the tools as shortening the distance between an intent and a diff you can check.
Other comparisons.
Each one gives a verdict first, then goes station by station.
Does this page answer your questions?
One tap, no sign-in.
What are you here for?
What is holding you back?
How does it read?
Thanks. If you left an email, Michał will reply.
Answering your questions?
One tap each. Close whenever you like.
no sign-in
Pick one. Learn the workflow once.
The guides cover Cursor, Claude Code and Codex side by side, in the same order, so a move to the other tool later costs you little.
$59.99 once for 12 months of access. Nothing renews on its own. 30-day refund on a first purchase.
Michał replies, not a bot. The address is used for this reply only.