The AI Coding Tools Landscape

The AI Coding Tools Landscape

The AI coding tools market is exploding. New tools launch every week, existing tools add agent capabilities overnight, and the categories themselves keep shifting. It is genuinely overwhelming -- even for people who work in this space full-time.

This lesson cuts through the noise. I am going to give you a comprehensive, opinionated map of every major AI coding tool in 2025-2026, organized by category, with honest assessments of when to use each one. Bookmark this lesson -- you will come back to it.

The Four Categories

AI coding tools fall into four distinct categories based on where and how they run:

+-----------------------------------------------------+
|               AI CODING TOOLS                        |
|                                                      |
|  +-------------+  +--------------+                   |
|  |  CLI AGENTS  |  |  IDE AGENTS  |                  |
|  |  Terminal    |  |  Editor      |                   |
|  |  based       |  |  integrated  |                  |
|  +-------------+  +--------------+                   |
|                                                      |
|  +--------------+  +--------------+                  |
|  | CLOUD AGENTS |  |  OPEN SOURCE |                  |
|  |  Remote      |  |  Community   |                  |
|  |  sandboxed   |  |  driven      |                  |
|  +--------------+  +--------------+                  |
+-----------------------------------------------------+

CLI Agents run in your terminal. They have direct access to your filesystem and can run any command. Maximum power, maximum flexibility, maximum responsibility.

IDE Agents run inside your code editor (VS Code, JetBrains, etc.). They integrate with your editing workflow and provide visual diffs, inline suggestions, and chat panels.

Cloud Agents run on remote servers in sandboxed environments. You give them a task and they work independently, often on a clone of your repository.

Open-Source Agents are community-built tools you can run locally, customize, and extend. They often work as VS Code extensions or CLI tools.

The Complete Comparison

Here is every major tool you need to know about:

CLI Agents

Claude Code (Anthropic)

• Model: Claude Opus 4, Sonnet 4 (configurable)
• Strengths: Most capable agent loop, deep codebase understanding, skills system, hooks, MCP integration, subagents, multi-surface (terminal, VS Code, JetBrains, web, mobile)
• Best for: Complex refactors, full-stack features, codebase migrations, CI/CD automation
• Revenue: $2.5B+ ARR (fastest-growing developer tool ever)
• Pricing: Included with Claude Pro ($20/mo), Team ($30/mo), Max ($100-200/mo), or API usage

Gemini CLI (Google)

• Model: Gemini 2.5 Pro, Gemini 2.5 Flash
• Strengths: Massive 1M+ token context window, strong multimodal (can analyze screenshots), free tier generous
• Best for: Large monorepos, projects with extensive documentation, multi-file analysis
• Pricing: Free tier (60 req/min), paid via Google AI Studio

Aider (Open Source)

• Model: Any (Claude, GPT, Gemini, local models)
• Strengths: Model-agnostic, git-native (auto-commits), lean and fast, excellent for pair programming
• Best for: Developers who want model flexibility and git integration
• Pricing: Free (bring your own API key)

IDE Agents

Cursor (Anysphere)

• Model: Claude, GPT, custom fine-tuned models
• Strengths: Best-in-class IDE experience, agent mode with multi-file editing, fast tab completion, composer for complex tasks
• Best for: Developers who live in their editor and want AI deeply integrated into the editing flow
• Pricing: Free tier, Pro $20/mo, Business $40/mo

GitHub Copilot (Microsoft/GitHub)

• Model: GPT-4.1, Claude Sonnet 4, Gemini (configurable)
• Strengths: Deepest VS Code integration, agent mode, huge ecosystem, coding standards, workspace indexing
• Best for: Teams already on GitHub, enterprise environments, VS Code power users
• Pricing: Free tier, Individual $10/mo, Business $19/mo, Enterprise $39/mo

Windsurf (OpenAI, acquired from Codeium)

• Model: GPT, Claude, custom models
• Strengths: "Cascade" multi-step agentic flows, good inline completions, strong free tier
• Best for: Developers who want a Cursor alternative with different UX
• Pricing: Free tier, Pro $15/mo

JetBrains AI Assistant (JetBrains)

• Model: Multiple providers
• Strengths: Native JetBrains IDE integration, understands project structure deeply, refactoring-aware
• Best for: JetBrains IDE users (IntelliJ, PyCharm, WebStorm, etc.)
• Pricing: Included with JetBrains All Products ($289/yr) or separate subscription

Cloud Agents

OpenAI Codex (OpenAI)

• Model: codex-1 (o3-based), GPT-5.3
• Strengths: True cloud sandbox, parallel task execution, fully autonomous, strong on SWE-bench
• Best for: Background task execution, processing multiple issues in parallel, teams wanting async AI coding
• Pricing: Included with ChatGPT Pro ($200/mo) or Team/Enterprise plans

Devin (Cognition AI)

• Model: Proprietary + Claude/GPT
• Strengths: Full development environment in the cloud, browser access, can deploy and test
• Best for: Teams wanting a fully autonomous "AI developer" for well-defined tasks
• Pricing: Team plan $500/mo

Amazon Q Developer (AWS)

• Model: Custom Amazon models
• Strengths: Deep AWS integration, infrastructure-as-code, security scanning, .NET migration
• Best for: AWS-heavy teams, Java/.NET enterprise environments
• Pricing: Free tier, Pro $19/mo

Open Source

Cline (VS Code Extension)

• Model: Any (Claude, GPT, Gemini, local)
• Strengths: Full agent capabilities in VS Code, model-agnostic, transparent tool use, MCP support
• Best for: Developers who want agent capabilities with full control and transparency
• Pricing: Free (bring your own API key)

Roo Code (Fork of Cline)

• Model: Any
• Strengths: Enhanced Cline with custom modes, boomerang orchestration, community-driven
• Best for: Cline users wanting more customization
• Pricing: Free (bring your own API key)

Continue (VS Code/JetBrains)

• Model: Any
• Strengths: Open-source, highly customizable, supports local models, privacy-focused
• Best for: Teams wanting full control over their AI coding setup
• Pricing: Free

The Decision Matrix

Choosing the right tool depends on your situation. Use this matrix:

Pricing Comparison (as of early 2026)

The Adoption Landscape

Here is what the market actually looks like in early 2026:

• GitHub Copilot still has the largest user base (~15M+ developers) due to first-mover advantage and GitHub integration
• Claude Code has the fastest growth and highest revenue per user, driven by its agent capabilities
• Cursor captured the "AI-native IDE" category and forced VS Code to accelerate AI features
• OpenAI Codex is gaining traction in enterprise for async agent workflows
• Gemini CLI is growing among Google Cloud/Android developers
• Open-source tools (Cline, Aider) have passionate communities and serve developers who want control

What This Course Focuses On

While we will touch on many tools throughout this course, we go deepest on:

1. Claude Code -- the most capable CLI agent with the richest feature set (skills, hooks, MCP, subagents)
2. The Agent Skills standard -- the open standard that works across 31+ tools
3. Cross-tool patterns -- workflows and mental models that apply regardless of which tool you use

The skills and patterns you learn here transfer across tools. If you master agent-directed development with Claude Code, you can apply the same techniques to Cursor, Copilot, Codex, or whatever comes next.