Integrated Development Environments (IDEs) and code editors in 2025 for AI!

It is crazy out there, there are many options for building code using AI that it is almost impossible to comprehend. So, today I will try to give some light to all of you, and also me. Lets start with the options you have when you are coding and you want to use something to help you.

There are now basically 2 options: (I am focusing on dedicated software development environment)

1. You can use a traditional Integrated Development Environment (IDE), like Visual Studio Code or Intellij. A piece of software you will install in your computer and you will code inside it. They will provide you with plugins or integrations that you install and then you can use ChatGPT, Claude or other LLMs.

   1. A small note here, you can always use your IDE without plugins or integrations and just copy paste your code in ChatGPT or Google Gemini :-)

2. You can use the so-called Agentic IDEs, they are modern and they have great integration with LLMs and Agents by default, they understand, plan and execute like a junior engineer. They usually offer you a plan and that plan includes several models that you can use while developing.

Ok, now that you know the options, let me show you examples of each one. Probably I will miss one or two but well the idea is to give you a general idea and not to be perfect.

1. Traditional Integrated Development Environments (IDE)

1.1 Visual Studio Code (VS Code): https://code.visualstudio.com/

Offer the following AI Integrations & Models: (They offer some free quota but usually you need to pay for a plan)

• Amazon Q Developer: (Amazon's proprietary LLMs)

  • https://aws.amazon.com/q/developer/

• GitHub Copilot: (OpenAI's LLMs like GPT-4o, o3-mini)

  • https://github.com/features/copilot/

• Google Gemini Code Assist: (Google's Gemini 2.5, Gemini 1.5 Pro/Flash)

  • https://cloud.google.com/gemini/docs/codeassist/

• Claude Code for VSCode (Anthropic)

  • https://docs.anthropic.com/en/docs/claude-code/ide-integrations

• JetBrains AI Assistant (via plugin): (JetBrains Mellum LLM; integrates with OpenAI GPT series, Anthropic Claude series, Google Gemini series; supports local LLMs via Ollama/LM Studio)

  • https://www.jetbrains.com/ai-assistant/

• Tabnine: (Tabnine's proprietary Universal AI models; Tabnine Chat can use Claude, GPT, Gemini, Llama, Mistral)

  • https://www.tabnine.com/install/vs-code/

• Qodo (called Codium in the past): (Proprietary AI models, uses Retrieval-Augmented Generation for codebase awareness)

  • https://www.qodo.ai/

  • https://marketplace.visualstudio.com/items?itemName=Codium.codium

• CodeGPT: (Connects to OpenAI, Microsoft, Google, Anthropic, Cohere, Mistral, Hugging Face models)

  • https://codegpt.co/

  • https://marketplace.visualstudio.com/items?itemName=DanielSanMedium.dscodegpt

• Windsurf (via plugin): (Windsurf's internal models, can leverage others like Claude Sonnet, it has collaborative agent called Cascade)

  • https://windsurf.com/

1.2 JetBrains IDEs (IntelliJ IDEA, PyCharm, WebStorm, etc.): https://www.jetbrains.com/

• Same as Visual Studio Code PLUS:

  • Native JetBrains AI Assistant: Deep integration with JetBrains' proprietary Mellum LLM and robust support for external models (OpenAI, Anthropic, Google) and local LLMs. Includes unique features like AI quick-fixes, cross-language conversion, and rename refactorings.

1.3 Eclipse: https://www.eclipse.org/

• Amazon Q Developer (Preview): (Amazon's proprietary LLMs)

• GitHub Copilot: (OpenAI's LLMs)

• Tabnine: (Tabnine's proprietary Universal AI models)

1.4 Vim / Neovim: https://www.vim.org/ and https://neovim.io/

• GitHub Copilot: (OpenAI's LLMs, via copilot.vim / copilot.lua plugins)

• Tabnine: (Tabnine's proprietary Universal AI models)

• Avante.nvim (Neovim specific): (Connects to various LLM providers via APIs: OpenAI, Anthropic, Google)

1.5 Sublime Text:

https://www.sublimetext.com/

• Tabnine: (Tabnine's proprietary Universal AI models)

• AskCodi: (Likely leverages various major LLMs)

  • https://www.askcodi.com/

2. Agentic IDEs

The list is short but they are growing:

2.1 Cursor: https://cursor.com/

• Models: Access to various AI models: OpenAI (GPT series), Anthropic (Claude series), Google (Gemini series). Supports self-hosted LLMs.

2.2 Windsurf: https://windsurf.com/

• Models: Windsurf utilizes its internal proprietary models (like SWE-1 Lite and SWE-1 for "Cascade Agent" and commands), and also provides access to leading third-party models such as GPT-4.1, GPT-4o, Claude Sonnet (including 3.5/3.7 versions), Deepseek R1, and Gemini 2.0 Flash (and other Gemini 2.5 models).

2.3 Kiro.dev: https://kiro.dev/

• Models: Primarily powered by Anthropic (Claude Sonnet 4.0 and 3.7), with plans for more models.

  • Key Capabilities: This is an agentic IDE (built on Code OSS/VS Code fork) with a unique "spec-driven development" approach. It helps define structured requirements, design, and tasks before coding, reducing "vibe coding" issues. Features include AI agents for tasks, "hooks" for automating repetitive workflows (like generating tests/docs on save), project steering, and deep codebase understanding via Model Context Protocol (MCP) servers.

2.4 OpenAI Codex (2025): https://openai.com/index/introducing-codex/

• Key Capabilities:

  • Autonomous task execution: writes features, fixes bugs, runs tests/linting, generates PRs

  • Available as both a ChatGPT sidebar agent (cloud sandbox) and a local CLI tool (codex npm package)

  • Runs in isolated environments, sandboxing code with approval control modes (suggest, auto-edit, full-auto)

  • Supports multimodal input (text, screenshots, diagrams) in CLI

So here they are, in a future post I might analyze the price and advantages of each one. But at least for the moment you have an idea of what you and your team can use for development. I left out many AI Tools most of them Narrow, Task-Specific AI Agents, Frameworks for Building Agents and General-Purpose Autonomous Agents (not focused on coding), because I wanted to focus on what you might actually find in your development team.

Happy coding!