Best AI tools for vibe coding

Day to day coding in an editor built around the model.

AI code editor for writing and editing a project from chat and tab completion.

Long jobs on a real repo: refactors, tests, and pull requests.

Anthropic coding agent that reads a repo, edits files, and runs commands.

Teams already on GitHub who want AI in the editor they use now.

GitHub coding assistant in the editor and on GitHub, from completions to pull requests.

An open agent when you want to choose the model yourself.

Open source coding agent for the terminal, a desktop app, and an editor. You pick the model.

The same OpenAI agent in chat, the terminal, and the cloud.

OpenAI coding agent in ChatGPT, an editor extension, and the terminal.

An editor plus a board of local and cloud agents.

The editor formerly called Windsurf. A full IDE plus a board for local and cloud coding agents.

Google agent harness, including a no-charge start.

Google agent-first coding platform with a desktop app, an IDE, and a CLI.

Spec first, then agents, especially if you work on AWS.

Amazon agentic IDE and CLI that turns a prompt into a spec, then implements it.

An open agent that asks before it acts.

Open source coding agent for editors and the terminal. You choose the model and approve steps.

A coding agent inside JetBrains IDEs and the terminal.

JetBrains coding agent for the terminal and JetBrains IDEs. You can bring your own model key.

Git commits from a terminal pair programmer.

Terminal pair programmer that maps the repo and commits its edits with git.

A running web app from a prompt, in the browser.

Browser app from StackBlitz that builds a running web app from a prompt.

UI and full-stack web apps you will deploy on Vercel.

Vercel prompt-to-app tool for UI and full-stack projects, with deploy built in.

A small product or internal tool from a description.

Builds a full-stack app from a description, then hosts it.

Building in the browser, with the app running in the cloud.

Browser IDE with an agent that can build and run the project in the cloud.

A local coding model you can run on your own machine.

Recursive's open-weight coding model for local and offline use, with about 40 billion total parameters.

A coding-capable model you call from MiniMax Code or the API.

MiniMax multimodal model for text, image, and video input, with a 1 million token context.

A small asset engine a coding agent can drive.

MIT-licensed engine that draws 2.5D game assets from TypeScript code and exports sprite sheets.

Kimi Code and the Kimi API, while the weights are still unreleased.

Moonshot's first K3.1 model, described as 3T-class, with a 1 million token context. Weights are not downloadable yet.

A dense 33B coding model you can download.

A 33B dense open-weight model for agentic coding and office tasks, under Apache 2.0.

A free CLI for Gemini models inside a shell or coding agent.

Google's experimental CLI for Gemini models, media generation, and Managed Agents. The CLI is free. API use is billed.

A Codex model on a Plus, Pro, Enterprise, or Edu plan.

OpenAI model for computer use, visual reasoning, and writing. Available in ChatGPT Work, Codex, and the API as gpt-6.1-terra.

A long-context coding model, once early access exists.

Mirendil's first model, pretrained from scratch, with a context window of up to 10 million tokens. Access is early-access only.

A frontier Gemini model in Antigravity or the Gemini API.

Google's frontier Gemini model, in the Gemini API, the Gemini app, and Google Antigravity.