AgentsPick an agent for coding
Start from where you write code and how much you want to hand off. Every coding agent we profile is here, grouped so you can scan it in one screen.
Inside the app where you write code
Tools like VS Code, Cursor or JetBrains, where developers write and edit code.
GitHub Copilot
Choose it if: You want the cheapest paid entry ($10/mo), work in almost any editor, and want to hand GitHub issues to an agent.
Free $0 (limited chat and agent use); Pro $10/mo; Pro+ $39/mo; Max $100/mo; Business $19/seat/mo; Enterprise $39/seat/mo. Chat, agents, code review and CLI use GitHub AI Credits (1 credit = $0.01), and each plan includes a monthly allowance. Code completions don't use credits on paid plans.
Cursor
Choose it if: You want an AI-first editor every day.
Free; Plus $8/mo; Premium $30/mo; Super $100/mo; Ultra $200/mo; Teams $30, $100 or $200 per seat. Legacy Pro ($20) plans continue for existing subscribers.
Devin
Choose it if: You want Windsurf (now Devin Desktop) or a fully autonomous cloud engineer for well-scoped tickets.
Free $0; Pro $20/mo; Max $200/mo; Teams $80/mo plus $40/mo per full dev seat (up to 200 users); Enterprise custom. Existing Windsurf plans carried over unchanged.
Kiro & Amazon Q Developer
Choose it if: You build on AWS or want a spec and plan before any code.
Kiro: Free (50 credits); Pro $20/user/mo; Pro+ $40; Pro Max $100; Power $200. Extra credits cost $0.04 each. Amazon Q Developer: Free tier (50 agentic requests a month); Pro $19/user/mo.
Junie
Choose it if: You already use IntelliJ, PyCharm or another JetBrains IDE.
Sold through JetBrains AI plans: AI Free (3 AI credits per 30 days), AI Pro $10 (10 credits), AI Ultimate $30 (35 credits). One credit equals $1 of usage.
Zed
Choose it if: You want a very fast editor with agents built in.
Personal $0 forever; Pro $10/mo, which includes $5 of tokens, with usage beyond that at API list price +10%; Business $30/seat/mo. There's a 14-day Pro trial with no card required.
Kilo Code
Choose it if: You want an open-source VS Code agent with pay-as-you-go model access.
Free and open source. Model usage is billed separately. Teams and Enterprise plans are on the pricing page.
Type-a-command and cloud coding agents from the AI labs
You type instructions in a text window on your computer (the terminal), or hand a task to an agent that works on its own cloud computer.
Claude Code
Choose it if: You work in the terminal (the text window where you type commands) and want an agent that edits and tests across a whole project.
Included with Claude Pro ($20/mo, or $17/mo billed annually) and Max (from $100/mo), and shares the plan's usage limits. Team Standard seats are $20/seat/mo annually ($25 monthly). You can also pay per token through the API.
OpenAI Codex
Choose it if: You already pay for ChatGPT and want coding agents included.
Included in ChatGPT Free and Go (limited trial access), Plus, Pro, Business and Enterprise. Pro comes in $100 and $200 tiers, and Business in $20 and $100 tiers, per OpenAI's Codex pricing page. Usage per 5-hour window depends on the model. You can also pay through an API key.
Gemini CLI & Jules
Choose it if: You want a free terminal agent, or background GitHub tasks with Jules.
Gemini CLI is free with a Google account: up to 1,000 requests a day (Google AI Pro 1,500, Ultra 2,000). Jules allows 15 tasks a day free, 100 on Pro and 300 on Ultra. Google AI Pro is $19.99/mo.
Grok Build
Choose it if: You subscribe to Grok and want xAI's terminal agent.
Grok Build in the Grok apps is on every plan, per xAI (Aug 19, 2026). Grok plans are sold as The X Subscription: Free; Plus $8/mo; Premium $30/mo; Super $100/mo; Ultra $200/mo.
Le Chat & Vibe
Choose it if: You want Mistral's Vibe coding agent from a European lab.
Free; Team $24.99/user/mo. Students get Pro for $14.99/mo for up to 12 months. The standard Pro price is shown on Mistral's pricing page, but we couldn't read it reliably, so check it there.
Qwen Code
Choose it if: You want Alibaba's open-weight Qwen coder models in a terminal agent.
Open source. Costs depend on the model provider you connect. Check the docs for current sign-in options.
Kimi Code
Choose it if: You want Moonshot's Kimi models as a low-cost coding plan.
Sold through Kimi Code Plans. We couldn't verify the prices on the official page today, so check kimi.com/code.
GLM Coding Plan
Choose it if: You want a low-cost plan that plugs GLM into Claude Code, Cline or OpenCode.
From $18/mo (Lite). Pro and Max are higher tiers.
DeepSeek
Choose it if: You want very cheap API models to power an agent you already use.
API priced per million tokens, with cheaper off-peak rates and cache-hit discounts. See the official pricing table.
Free, open-source agents that can run on your own computer
Anyone can inspect the code, and you can pair them with a model running on your own machine so your code stays private.
Cline
Choose it if: You want an open-source agent in VS Code with any model, including local ones.
Free and open source for individuals. You pay only for model usage, either with your own API keys or through the Cline provider. Enterprise is custom.
Continue
Choose it if: You want VS Code or JetBrains with a local model from Ollama or LM Studio.
Free and open source (Apache 2.0). With local models through Ollama there is no usage fee.
Aider
Choose it if: You want git-native, open-source pair programming in the terminal.
Free and open source (Apache-2.0). You pay only for the model API you connect, or nothing if you use a local model.
OpenCode
Choose it if: You want a model-agnostic open-source agent across 75+ providers.
Free and open source. You pay only for the models you connect.
Build an app without writing code
Describe the app you want in plain English and the agent builds and hosts it.
Replit Agent
Choose it if: You are not a developer and want an app built and hosted from a description.
Core $20/mo ($18/mo billed annually), which includes $20 toward more powerful models. The next tier is $100/mo ($90 annually) with $100 toward models and up to 15 collaborators. Enterprise is custom.
Local and private option
Private or offline? Pair Aider, Cline, Continue or OpenCode with a Qwen coder model in Ollama or LM Studio so code never leaves your machine. See the local vs cloud comparison, the model picker for what fits your GPU, and the local agents guide (/learn/local-agents) for VS Code with Ollama or Continue. Open the model picker →
Guardrails
- Agents can run commands. Keep approvals on for anything destructive.
- Review every diff before merging.