Best AI coding agents for the terminal (CLI)
CLI coding agents live where your build tools already are. They read the repo, edit files and run commands from the terminal, which makes them easy to script and to pair with any editor.
Ranked by score on our 100-point rubric (how we score). Scores are initial editorial estimates from public information, not hands-on tests. Pricing is labeled verified (checked on the vendor's official pricing page, with date) or unverified.
Quick picks
| # | Tool | Tier | Score |
|---|---|---|---|
| 1 | Claude Code Agentic coding in your terminal and IDE | S | 89 |
| 2 | OpenAI Codex Cloud and CLI coding agent from OpenAI | S | 88 |
| 3 | Gemini CLI Open-source terminal agent powered by Gemini | A | 81 |
| 4 | OpenCode Open-source terminal coding agent | A | 77 |
| 5 | Aider Terminal pair programmer with git integration | A | 73 |
The ranking
-
Anthropic's CLI coding agent that reads your repo, edits files, runs commands and integrates with MCP.
- Best for
- You want the strongest autonomous agent for real multi-file work.
- Skip it if
- You want a free tool or a GUI-only workflow.
- Pricing
- Included with Claude Pro ($20/mo, or $17/mo annually) and Max (from $100/mo); not included in the Claude Free plan ✓ verified on the official pricing page, Oct 18, 2026
- Platforms
- CLI · Mac · Win · Linux · IDE extension · MCP
-
OpenAI's coding agent available as an open-source CLI, IDE extension and cloud task runner.
- Best for
- You have ChatGPT and want parallel cloud tasks plus a local CLI.
- Skip it if
- You are not in the OpenAI ecosystem.
- Pricing
- Included with ChatGPT Plus/Pro/Business; API usage also available unverified editorial summary, not yet confirmed on an official pricing page
- Platforms
- CLI · web · IDE extension · Mac · Linux · Win
-
Google's open-source CLI agent with a large context window, MCP support and a free usage tier.
- Best for
- You want a capable free terminal agent.
- Skip it if
- You need the very best code quality on hard refactors.
- Pricing
- Free tier with a personal Google account; paid via API keys unverified editorial summary, not yet confirmed on an official pricing page
- Platforms
- CLI · MCP
-
Terminal-first open-source agent with a polished TUI that works with 75+ model providers.
- Best for
- You want a Claude Code-style experience with any model.
- Skip it if
- You want a vendor-supported product.
- Pricing
- Free and open source; pay your model provider unverified editorial summary, not yet confirmed on an official pricing page
- Platforms
- CLI
-
Open-source CLI that edits code in your repo and auto-commits, working with most LLMs.
- Best for
- You like the terminal and clean git commits per change.
- Skip it if
- You want a GUI or a fully autonomous agent.
- Pricing
- Free and open source; pay your model provider unverified editorial summary, not yet confirmed on an official pricing page
- Platforms
- CLI
Head-to-head comparisons
FAQ
What is the best AI tool for terminal (CLI) coding agents?
Claude Code ranks first with 89/100 (S-tier). Agentic coding in your terminal and IDE. Use it when you want the strongest autonomous agent for real multi-file work. Next in our ranking: OpenAI Codex (88/100, S-tier) and Gemini CLI (81/100, A-tier).
Which AI tools for terminal (CLI) coding agents have a free tier?
In our pricing data, Gemini CLI (unverified), OpenCode (unverified) and Aider (unverified) list a free tier. "Unverified" means it is an editorial summary we have not yet confirmed on the vendor's official pricing page, so check the vendor before relying on it.
Are there open-source options for terminal (CLI) coding agents?
Yes. OpenAI Codex, Gemini CLI, OpenCode and Aider are listed as open source or open weights. Check each project's license before commercial use.
How are these tools ranked?
Each tool is scored out of 100 on our public rubric: job fit (30), ease (15), value (15), momentum (15), openness and portability (10), trust and maturity (10) and a builder/gamer bonus (5). S-tier is 85+, A-tier 70–84, B-tier 55–69 and C-tier below 55. Current scores are initial editorial estimates from public information, not hands-on tests, and sponsors can't buy points.