mrkeyoor.com_
Tue 10 Nov 03:47 UTC

Best Open-Source AI Coding Agents (2026) — measured, not ranked by stars

We cloned and ran every agent on this page in a clean container; the numbers are ours. Each one got the same treatment: a fresh checkout, the project’s own install command, then its own build and test commands, on a 3-CPU, 8 GB sandbox with a 900-second cap per step. Where a project exposes no build or test target, we say so instead of guessing. The order below is our usability score first, stars only as a tie-break — which is why a 27k-star agent outranks a 119k-star one.

Twelve terminal and editor agents, each cloned and run on our own machine — install time, dependency count, build and test results.

01

crush

charmbracelet/crush · ★ 29k ▲67/7d · Go

Choose Crush if you want a terminal coding agent with model switching, language-server context and MCP connections, and can review executable project configuration before launch. Crush lets you switch models mid-session while preserving context. Our September 30, 2026 sandbox run checked repository command outcomes at commit 89c3c4a; it did not evaluate model responses or completed coding tasks.

Measured sandbox golang:1.24-bookworm · 3 cpu / 8g · install 166s · 446 deps · build 211s · tests 118/118 go test · 167,608 LOC · 642s total

SetupDocsCommunityMaturity

Verdict Evaluate Crush if model choice and terminal integration fit your workflow. Before adopting it, review executable configuration, keep tool approvals enabled and try one representative coding task with your own provider and terminal.

Pick it if Developers who prefer a terminal interface and want to choose their model provider.

Full review →
02

codex

openai/codex · ★ 134k ▲1.2k/7d · Rust

Choose Codex CLI if you want OpenAI's terminal coding agent and its documented ChatGPT or API-key access fits your workflow. Codex CLI is a coding agent from OpenAI that runs locally on your computer. Our run checked repository installation, not an authenticated coding task, so it cannot establish how well the agent would handle your project.

Measured sandbox lab-node:22 · 3 cpu / 8g · install 27s · 1 deps · 170 MB · no build target · no test target · 2,070,775 LOC · 66s total

SetupDocsCommunityMaturity

Verdict Codex CLI is worth trying when you want OpenAI's terminal agent and can use its documented account access. Start with a released installation and verify a small task in a disposable repository.

Pick it if Developers who want the terminal interface rather than Codex Web, the IDE integration, or the desktop app.

Full review →
03

goose

block/goose · ★ 56k ▲179/7d · Rust

Goose is a local AI agent available as a desktop app, command-line tool, and embeddable API. It can edit code, run shell tools, automate other work, and connect to outside capabilities through MCP extensions while letting you choose the model provider.

Measured sandbox lab-rust:1 · 3 cpu / 12g · install 66s · 1,389 deps · build FAILED after 900s (900s cap) · tests FAILED after 900s (900s cap) · 429,650 LOC · 1,911s total

SetupDocsCommunityMaturity

Verdict Our Goose run installed 1,389 packages in 66 seconds, but both build and tests reached their 900-second caps while Rust was still compiling, so source contributors should expect a slow first feedback loop. The packaged app is the better route for people who want a provider-flexible local agent with MCP tools.

Pick it if Developers who want one local agent across a desktop app, terminal, and API.

Full review →
04

qwen-code

QwenLM/qwen-code · ★ 29k ▲130/7d · TypeScript

Qwen Code is an open-source coding agent for the terminal, with documented editor, desktop, browser and chat interfaces. Its README lists OpenAI, Anthropic, Gemini and Qwen APIs, plus local endpoints. These are documented capabilities; our recorded source-checkout run did not verify model compatibility or coding quality.

Measured sandbox lab-node:22 · 3 cpu / 8g · install 816s · 1 deps · 1,398 MB · build 361s · tests FAILED after 900s (900s cap) · 4,383,804 LOC · 2,110s total

SetupDocsCommunityMaturity

Verdict Qwen Code merits a trial when provider choice matters. The recorded source setup and build completed, while the test command reached its time limit without a completed suite result.

Pick it if Developers evaluating several model providers through one coding client.

Full review →
05

opencode

sst/opencode · ★ 226k ▲1.1k/7d · TypeScript

OpenCode is an open-source coding agent that works in a terminal and, in beta, as a desktop app. It gives developers a read-only planning mode and a full-access build mode, so the same tool can inspect a codebase or change it under different permission rules.

Measured sandbox lab-node:22 · 3 cpu / 8g · install FAILED after 900s (timeout) · deps: not measured · build: not measured · tests: not measured · 688,859 LOC · 915s total

SetupDocsCommunityMaturity

Verdict Our OpenCode install reached the 900-second cap without finishing, so contributors should prove the Bun workspace on their own machine before committing to it. The product is still a strong candidate for developers who want open source, model choice, and a clear split between planning and editing.

Pick it if Developers who want an open-source coding agent in the terminal.

Full review →
06

cline

cline/cline · ★ 71k ▲153/7d · TypeScript

Choose Cline if you want a coding agent with explicit approval controls and a choice of model providers. Its source build completed in our sandbox, but the test command did not pass; that run does not establish successful coding tasks or deployment readiness.

Measured sandbox lab-node:22 · 3 cpu / 8g · install 20s · 2,446 deps · 3,897 MB · build 96s · tests 150/150 vitest · 857,522 LOC · 451s total

SetupDocsCommunityMaturity

Verdict Trial Cline for a small, reviewable coding task when approval controls and provider choice match your needs. Our October 14, 2026 run at commit 1932d5d supports a successful source build, but its test command failed and no agent workload was verified.

Pick it if Developers who want to review file edits and terminal commands before allowing them.

Full review →
07

aider

Aider-AI/aider · ★ 50k ▲93/7d · Python

Choose Aider for a supervised terminal coding workflow if you want to review model edits through Git. Aider automatically commits changes with sensible commit messages. Our run on September 30, 2026, at commit 5dc9490 installed and built successfully, but its test command failed. That run does not establish the quality of model-generated patches or successful unattended operation.

Measured sandbox ghcr.io/astral-sh/uv:python3.12-bookworm · 3 cpu / 8g · install 154s · 171 deps · 772 MB · build 8s · tests 474/492 pytest · 42,721 LOC · 778s total

SetupDocsCommunityMaturity

Verdict Aider merits a supervised trial if terminal interaction and Git review fit your work. Our September 30, 2026 run at commit 5dc9490 supports successful installation and a zero-exit build step, followed by a failed test command.

Pick it if Developers who want terminal sessions and Git-based review of AI edits.

Full review →
08

gemini-cli

google-gemini/gemini-cli · ★ 109k ▲17/7d · TypeScript

Check account eligibility before installing Gemini CLI. Google's current README describes it as an open-source terminal agent maintained for enterprise Gemini Code Assist workflows on Google Cloud. Our October 1, 2026 source run installed and built successfully, but testing timed out. Authenticated coding, model quality, deployment, and the documented sandbox providers were not verified.

Measured sandbox lab-node:22 · 3 cpu / 8g · install 46s · 1,336 deps · 798 MB · build 95s · tests 7,187/7,201 vitest · 703,120 LOC · 1,055s total

SetupDocsCommunityMaturity

Verdict Consider Gemini CLI when its current enterprise Code Assist scope matches your account and workflow. The October 1, 2026 run supports installation and building of the recorded source revision, but leaves testing unfinished and authenticated coding unverified.

Pick it if Organizations using Gemini Code Assist Standard or Enterprise, or Code Assist for GitHub on Google Cloud.

Full review →
09

continue

continuedev/continue · ★ 36k ▲43/7d · TypeScript

Continue is an open-source coding assistant for the terminal, VS Code, and JetBrains editors. It connects your coding workflow to language models, but the README now calls 2.0.0 the final release and says the repository is no longer actively maintained.

Measured sandbox lab-node:22 · 3 cpu / 8g · install 35s · 319 deps · 107 MB · no build target · no test target · 331,015 LOC · 54s total

SetupDocsCommunityMaturity

Verdict Our Continue install succeeded in 35 seconds, but npm audit reported 12 known vulnerabilities and our run verified neither a build nor tests. Combined with the README's maintenance notice, that makes Continue a deliberate fork or migration decision for existing users.

Pick it if Existing Continue users deciding whether to keep a pinned client while arranging a replacement.

Full review →
10

OpenHands

All-Hands-AI/OpenHands · ★ 93k ▲395/7d · TypeScript

OpenHands Agent Canvas is a browser interface for managing coding agents across local, remote, and cloud backends. Its README describes support for OpenHands and external ACP agents, with a separate automation service for scheduled and event-triggered work. Our recorded run checked repository dependency installation, a build, and an unfinished test command; it did not verify a working agent deployment.

Measured sandbox lab-node:22 · 3 cpu / 8g · install 47s · 1,959 deps · 1,057 MB · build 24s · tests FAILED after 900s (900s cap) · 416,090 LOC · 980s total

SetupDocsCommunityMaturity

Verdict OpenHands Agent Canvas is worth evaluating when managing agents across several execution backends solves a specific workflow problem. Our October 14, 2026 run at commit 0e9ce19 completed dependency installation and a build, but left tests unfinished at the 900-second cutoff.

Pick it if Developers who need a shared browser interface for agents running on different machines.

Full review →
11

plandex

plandex-ai/plandex · ★ 16k ▲9/7d · Go

Plandex is a terminal coding agent for tasks that span multiple steps and files. Its README describes a cumulative diff that keeps generated changes separate from project files before application, plus plan history and branches. That review workflow is a reason to evaluate it; it does not establish filesystem isolation or correct edits.

Measured sandbox golang:1.24-bookworm · 3 cpu / 8g · install 35s · 360 deps · build 44s · tests passed in 10s (0 collected) · 81,449 LOC · 115s total

SetupDocsCommunityMaturity

Verdict Plandex merits a restricted evaluation if its plan-and-diff workflow fits your work and you can operate the local server. Our September 28, 2026 run of commit e2d7720 recorded a 35-second install step and a successful 44-second build step, but also an unsuccessful vet result and no demonstrated behavioral tests in the CLI command scope.

Pick it if Developers evaluating a planning workflow for changes across multiple files.

Full review →
12

kilocode

Kilo-Org/kilocode · ★ 28k ▲58/7d · TypeScript

Kilo Code is an open-source coding agent available in VS Code, JetBrains IDEs and a terminal, with hosted cloud and pull-request review options. Its README describes file editing, repository questions and command execution, with a choice of models.

Measured sandbox lab-node:22 · 3 cpu / 8g · install 80s · 1,996 deps · 2,629 MB · no build target · tests FAILED after 12s · 1,378,588 LOC · 105s total

SetupDocsCommunityMaturity

Verdict Our Kilo Code install recorded 1,996 dependencies and 2,629 MB, while the root test command refused to run tests. Try a published client if you want its IDE and terminal options, but budget separately for contributing to the source workspace.

Pick it if Developers evaluating a coding assistant for an IDE and a terminal.

Full review →

Also run, not ranked: Roo-Code installed 2,665 packages and passed build and tests, but the extension was shut down and the repository archived on 15 May 2026, so it is not a 2026 pick.

See also