AI Coding / Vibe Coding

AI Coding Tools, Models, and Workflow Guide

Start from your intent: pick a tool, choose a model, ship AI-built work safely, or learn vibe coding without losing control.

Choose a toolChoose a modelShip AI-built workVibe coding safely

Last reviewed: 2026-07-17. Tool and model listings show per-item review dates; rankings use dated snapshots.

Start from your intent

Four clear paths into the hub. Each path states who it is for, what it does not replace, and where to go next.

I want to choose an AI coding tool

Pick by workflow: editor, terminal agent, coding agent, or cloud builder—then compare Cursor, Claude Code, and Codex.

Best for: Developers and teams selecting a primary AI coding setup.

Limits: Not a permanent #1 ranking. Product surfaces change; verify official docs.

Tool selection guideBrowse tools directory

I want to choose a coding model

Read dated leaderboard snapshots for cost, latency, and context—without treating scores as forever quality ranks.

Best for: Builders comparing models inside an existing tool workflow.

Limits: Public benchmarks are not a private coding exam for your repo.

Model methodology guideOpen model rankings

I already used AI to write a project

Verify diffs, builds, secrets, and release readiness with local checklists before public users arrive.

Best for: Anyone shipping AI-generated or AI-assisted changes.

Limits: Checklists do not upload your repo or auto-judge code quality.

Ship checklistWorkflow guide

I am a non-technical builder

Break ideas into slices, control permissions and data, accept work against criteria, and plan deploy plus rollback.

Best for: Founders, PMs, and builders using AI app builders or agents.

Limits: Does not replace security review for payments or sensitive data.

Vibe coding guideProject rules overview

Recommended AI coding tools

See more AI coding tools
AI coding IDE

Cursor

Cursor combines an editor, chat, agent runs, inline edits, and codebase context for day-to-day software development.

Best for: Developers and teams that want AI assistance inside the editor.

Watch outs: Evaluate rule files, repository indexing, pricing, and how much control the team needs over model behavior.

Read introductionLast reviewed: 2026-07-19Terminal agent

Claude Code

Claude Code is designed for agentic editing, explanation, debugging, and command-line development loops.

Best for: Engineers comfortable working from a terminal and reviewing diffs.

Watch outs: Check command permissions, repository boundaries, and how the team records agent decisions.

Read introductionLast reviewed: 2026-07-19Coding agent

Codex

Codex represents OpenAI's coding-agent workflow across app, terminal, editor, and multi-agent development surfaces.

Best for: Teams already building around OpenAI models and agent workflows.

Watch outs: Confirm the exact product surface, model availability, and integration path before adoption.

Read introductionLast reviewed: 2026-07-19Agentic IDE

Windsurf

Windsurf focuses on AI-assisted coding inside an IDE, with agent flows and codebase-aware development.

Best for: Developers comparing Cursor-style editor agents and team coding environments.

Watch outs: Review product changes, account migration, and roadmap stability when comparing it with other IDEs.

Read introductionLast reviewed: 2026-07-19AI IDE

Trae

Trae positions itself as an adaptive AI IDE for coding, builder workflows, multimodal input, and project-context assistance.

Best for: Teams evaluating newer AI IDEs and builder-style coding environments.

Watch outs: Check regional availability, platform support, model choices, and data handling before standardizing.

Read introductionLast reviewed: 2026-07-19Terminal agent

opencode

opencode focuses on terminal-based AI coding with a lightweight interface and model-provider choice.

Best for: Developers comparing newer terminal agents.

Watch outs: Validate provider setup, local permissions, and repository safeguards.

Read introductionLast reviewed: 2026-07-19

AI coding model rankings

See top 20 models
01

Anthropic - Claude Fable - 5

Still the highest Intelligence Index in the snapshot, with a 1M context window.

Click for model details
Context
1M
AA Index
60
Blended price
$7.70/1M
Speed
66 tok/s

Best for: Deep repository analysis, complex planning, and high-value agent tasks where latency is acceptable.

02

OpenAI - GPT Sol - 5.6 (max)

New GPT-5.6 Sol tier at the top of OpenAI's reasoning lineup in this snapshot.

Click for model details
Context
1M
AA Index
59
Blended price
$4.35/1M
Speed
52 tok/s

Best for: OpenAI-centered coding agents, long-context planning, and max-reasoning software tasks.

03

OpenAI - GPT Sol - 5.6 (xhigh)

High-reasoning GPT-5.6 Sol mode with 1M context and much lower latency than max.

Click for model details
Context
1M
AA Index
58
Blended price
$4.35/1M
Speed
52 tok/s

Best for: Daily agent coding, code review, and tool-heavy workflows that still need strong reasoning.

04

Kimi - Kimi - K3

Moonshot's new 2.8T MoE flagship with 1M+ context, strong coding scores, and low first-chunk latency.

Click for model details
Context
1.05M
AA Index
57
Blended price
$2.31/1M
Speed
62 tok/s

Best for: Long-horizon coding agents, frontend generation, regional API stacks, and cost-aware 1M context work.

05

OpenAI - GPT Sol - 5.6 (high)

Balanced GPT-5.6 Sol high tier with 1M context and responsive total response time.

Click for model details
Context
1M
AA Index
56
Blended price
$4.35/1M
Speed
47 tok/s

Best for: Interactive coding sessions, refactors, and product engineering loops inside OpenAI tools.

06

Anthropic - Claude Opus - 4.8 (max)

High-intelligence Claude Opus 4.8 with 1M context and competitive end-to-end latency.

Click for model details
Context
1M
AA Index
56
Blended price
$3.85/1M
Speed
56 tok/s

Best for: Large code migrations, multi-file edits, and coding agents that need strong reasoning.

Third-party plugins and skills

See more plugins

Reviewed repositories with at least 10,000 GitHub Stars are assigned to one primary capability and ordered by a dated snapshot. Secondary capabilities appear only on the detail page. Stars show attention, not an absolute quality or security score.

#01Coding & implementationCollection

Superpowers

A high-signal Claude Code extension project for skills, workflow upgrades, and agent coding habits.

Claude Code users who want battle-tested workflow extensions.GitHub Stars257.3K2026-07-19Read introductionLast reviewed: 2026-07-19
#02Coding & implementationCollection

ECC

A large engineering system with skills, rules, agents, memory, security, and research-first workflows for major coding agents.

Developers who want a comprehensive, configurable coding-agent setup across languages and tools.GitHub Stars231.1K2026-07-19Read introductionLast reviewed: 2026-07-19
#03Planning, requirements & specsSkill

Grill Me

Matt Pocock's viral /grill-me skill: one-question-at-a-time interviews that stress-test a plan before any code is written.

Developers who want to surface hidden assumptions in a design or feature plan before agentic coding starts.GitHub Stars177K2026-07-19Read introductionLast reviewed: 2026-07-19
#04Coding & implementationCollection

Anthropic Skills

Anthropic's public skills repository for Claude agents: reusable workflows, document skills, and reference patterns.

Teams standardizing Claude Code, Cowork, or API agents with checked-in skills.GitHub Stars162.6K2026-07-19Read introductionLast reviewed: 2026-07-19
#05Agent orchestration & automationPlugin

gstack

A Claude Code plugin with opinionated product, design, engineering-management, release, documentation, and QA roles.

Founders and small teams that want forceful reviews across the full product-development loop.GitHub Stars122.9K2026-07-19Read introductionLast reviewed: 2026-07-19
#06Planning, requirements & specsSpec tool

GitHub Spec Kit

GitHub's toolkit for turning product intent into specifications, plans, and implementation tasks.

Teams adopting a repeatable spec-driven workflow across different AI coding agents.GitHub Stars122.2K2026-07-19Read introductionLast reviewed: 2026-07-19

AI coding courses

Published course

AI Literacy for Developers

Practical AI basics for developer workflows: models, prompts, coding assistants, verification, privacy, and reliable AI-assisted work.

7 lessonsOpen course

Planned course series

Planned

AI Coding Tools in Practice

Project setup, rules, permissions, plugins, and review workflows for Cursor, Claude Code, Codex, and OpenCode.

Planned

Vibe Coding Product Practice

From idea and prototype to requirements, boundaries, launch checks, and responsible publishing.

Planned

AI Agent Engineering

Specs, subagents, MCP, tests, PR review, release checks, and reusable team workflows.

Content columns

Tools

Tool selection

Choose AI coding tools by workflow: editors, terminal agents, coding agents, and cloud builders.

Compare

Cursor vs Claude Code vs Codex

Compare environment, permissions, context, diff review, and team fit—without a permanent winner.

Models

Coding model rankings

Learn how to read dated model snapshots for cost, latency, context, and real project fit.

Models

Kimi K3 launch brief

Public specs, AA #4 placement, coding benchmarks, and API notes for Moonshot's new flagship (2026-07-17).

Workflow

Specs and agent workflow

Cover specs, task slices, plan-review loops, test feedback, PR flow, and team habits.

Vibe

Vibe coding

Teach builders how to describe ideas, inspect outputs, avoid demo traps, and ship responsibly.

Rules

Project rules and specs

Explain AGENTS.md, CLAUDE.md, Cursor Rules, MCP, specs, and verification gates.