You installed the tool, opened a project, and now you have questions. This page answers the ones developers actually ask — from “which model should I use?” to “why did I hit my message limit after ten prompts?”
Cursor is an AI-native IDE (fork of VS Code) with inline editing, tab completions, and agent mode built in. Claude Code is a CLI-first agent that lives in your terminal and works alongside any editor. Codex is OpenAI’s multi-surface agent — a web app, CLI, IDE extension, and cloud service with GitHub/Slack/Linear integrations.
Pick Cursor if you want everything in one IDE. Pick Claude Code if you live in the terminal. Pick Codex if you want cloud-based automation and multi-platform access.
| Aspect | Cursor | Claude Code | Codex |
|---|---|---|---|
| Interface | Full IDE (VS Code fork) | Terminal / CLI | App + CLI + IDE + Cloud |
| Primary model | Configurable (Opus 4.8, Composer 2.5, GPT-5.5, etc.) | Claude Opus 4.8 (Claude Fable 5 from v2.1.170) | GPT-5.5 |
| Context window | Model-dependent, up to 1M | 1M (Opus 4.8) | 400K via Codex (1M via API) |
| Sandboxing | Sandboxed terminals (macOS) | Docker optional | Git worktrees (default) |
| Cloud execution | Background Agent | Headless mode | Cloud agents (native) |
| MCP support | Yes | Yes | Yes |
| Skills support | Yes | Yes | Yes |
| Best for | IDE-first developers | Terminal-first developers | Multi-surface, automation |
| Starting price | $20/month (Pro) | $20/month (Pro) | Free tier + Pro |
For maximum capability (when velocity and quality outweigh cost): Claude Fable 5 — the new tier above Opus (June 2026), much better at complex refactorings, bug-fixing, and building apps from scratch; $10/$50 per 1M tokens (2x Opus 4.8). Included on paid Claude plans through June 22, 2026. Tip: as a default model it stays affordable because subagents still auto-run on cheaper tiers.
For complex coding tasks (default): Claude Opus 4.8 — top SWE-Bench scores, excellent agentic performance, strong reasoning.
For everyday work on a budget: Claude Sonnet 4.6 — excellent performance at roughly one-fifth the cost of Opus 4.8, plus a 1M token context window.
For Codex users: GPT-5.5 — the default model across all Codex and ChatGPT surfaces, with native computer-use and a 1M-token API context window (400K via Codex).
For speed-critical work in Cursor: Cursor Composer 2.5 — frontier-class coding model with MoE architecture at $0.50/$2.50 per 1M tokens. Roughly one-tenth the cost of the flagships and competitive on most benchmarks, though it trails GPT-5.5 on Terminal-Bench 2.0.
For extreme context or multimodal: Gemini 3.1 Pro — 1M token context, best image/video analysis, Deep Think mode.
See the full Model Comparison for detailed specs and pricing.
Individual Plans
Team Plans
Claude Fable 5: From June 9 through June 22, 2026, Fable 5 is included on Pro, Max, Team, and seat-based Enterprise plans at no extra cost; on June 23, 2026 it is removed from those plans and further use requires usage credits.
See the full Pricing Calculator for detailed cost scenarios.
Absolutely. Many developers use Cursor for interactive editing and tab completions, Claude Code for terminal-based autonomous tasks, and Codex for cloud-based automation and GitHub issue handling. The tools are complementary, not mutually exclusive. Your project configuration files (.cursor/rules/, CLAUDE.md, AGENTS.md) can coexist in the same repository.
All three tools offer privacy guarantees at paid tiers:
Always verify the current privacy policy for your tier. Free tiers may have different terms.
Open the AI chat panel (Cmd/Ctrl+I), then click the mode selector dropdown at the top. Choose “Agent” for autonomous file editing or “Ask” for read-only exploration. You can also press Cmd/Ctrl+. to toggle modes.
Settings > Models > API Keys. Enter your Anthropic, OpenAI, or Google key, click Verify, then select “Use my API key” in the model dropdown. Own API keys work for chat models but not Tab completion or specialized features.
.cursor/bugbot.yml in your repositorynpm install -g @anthropic-ai/claude-codeclaude --versionbrew install --cask claude-codeclaude --versionThen run claude in any project directory. You will be prompted to authenticate on first run.
These are not slash commands — they are keywords you type inside a normal prompt. Claude Code interprets phrases like think, think hard, think harder, and ultrathink as natural-language instructions that ask the model to allocate more reasoning before acting. Higher-intensity phrases request progressively deeper deliberation:
| Phrase in your prompt | Relative reasoning depth | Best for |
|---|---|---|
| (none) | Standard | Simple, well-scoped tasks |
think | More | Moderate complexity |
think hard / think harder | Deeper | Architecture decisions, tricky bugs |
ultrathink | Deepest | System design, gnarly debugging |
For example: “Refactor this auth middleware to support API keys. Think hard about backward compatibility.” Use deeper reasoning only when the task genuinely needs it — most coding tasks do fine with a plain prompt or a single think.
Granular approach (recommended): Configure allowed tools in your settings or use Claude Code’s permission system to pre-approve specific operations.
Full trust approach (use with caution):
claude --dangerously-skip-permissionsBest practice: Use --dangerously-skip-permissions only in sandboxed environments (Docker containers, CI pipelines).
Claude Code Pro allows roughly 10-40 messages per 5-hour rolling window (varies by complexity). Long conversations with big files consume more capacity. Strategies:
/compact to summarize long conversations/clear between unrelated tasksYes. Install the “Claude Code” extension from the VS Code marketplace. It launches a Claude Code terminal panel inside your editor. You get the full CLI experience without leaving your IDE.
Codex is available in four forms:
codex). Interactive and headless modes. Open-source.Every Codex task runs in an isolated Git worktree — a separate checkout of your repository. This means:
You do not need to manage worktrees manually. Codex creates and cleans them up automatically.
Codex can respond to external events automatically:
@codex or mention it in a PR commentEach automation spawns a cloud agent that executes the task and creates a PR or posts results back to the source.
| Mode | File reads | File writes | Command execution |
|---|---|---|---|
on-request | Yes | Asks first | Asks first |
on-failure | Yes | Yes | Auto-approves, asks on errors |
never | Yes | Yes | Yes (full autonomy) |
--full-auto flag | Yes | Yes | Yes (sandboxed) |
Start with on-request until you trust the agent’s output, then graduate to on-failure. Use --full-auto or approval_policy = "never" only with sandboxed environments or when you are confident in the task scope.
MCP (Model Context Protocol) servers extend what your AI agent can do by connecting it to external tools — databases, APIs, browsers, and more. You do not strictly need them, but they unlock powerful workflows. For example, the GitHub MCP lets the agent create PRs directly, and the Postgres MCP lets it query your database.
All three tools support MCP. Start with one or two essential servers and add more as needed.
MCP servers maintain a persistent connection and expose multiple tools. They are heavier to set up but more powerful for deep integrations (databases, browser automation, cloud services).
Agent Skills are lightweight, single-purpose augmentations installed via npx skills add <owner/repo>. They work across 35+ agents and are easier to share. Think of Skills as focused recipes and MCP servers as full integration platforms.
When both exist for a workflow, Skills are faster to set up; MCP servers offer deeper, persistent access.
Settings > Features > MCP > ”+ Add MCP Server”. Choose from the gallery or add a custom server configuration.
claude mcp add github -s user -- npx -y @modelcontextprotocol/server-githubOr add directly to .mcp.json in your project root.
codex mcp add github -- npx -y @modelcontextprotocol/server-githubOr add a [mcp_servers.github] table to ~/.codex/config.toml (project-scoped .codex/config.toml for trusted projects). The Codex IDE extension exposes this via MCP settings > Open config.toml.
A systematic approach for building complex features with AI:
think hard keyword or use extended reasoning for complex plansThis works identically across all three tools. The key is giving the AI a structured starting point rather than a vague request.
Use Agent mode (or equivalent autonomous mode) when:
Use Ask/Chat/Suggest mode when:
/clear in Claude Code, new chat in Cursor/cost, Codex dashboard