Best Coding agents — Open models

The ranking

1. ★ Editor’s pick

   #1KC

   Kilo Code

   Open source AI coding agent for VS Code, JetBrains, CLI and cloud, with 500+ models at provider rates.

   Best forOpen-weight hosted models plus fully local models through Ollama or LM Studio.

   Open models: Yes — Open-weight hosted models plus fully local models through Ollama or LM Studio.“You can also bring your own API keys for any provider, or run fully local models through Ollama or LM Studio when you need to keep code on your machine.”Frequently asked questions, "Which AI models does Kilo Code support?" · checked 2026-09-02

   Visit Kilo Code ↗Read the review →Compare Kilo CodeGet a quote →From Free
2. #2GO

   goose

   Open-source local AI agent — desktop app, CLI and ACP server — running on the LLM provider you bring.

   Best forDevs who want an open source coding agent that reuses an existing Claude or ChatGPT plan

   Open models: Yes — Open-weight models run locally through Ollama, LM Studio, Docker Model Runner, Ramalama or goose's own local-models command, and remotely through gateways such as Novita, OVHcloud and Cerebras.“Local model runner supporting Qwen, Llama, DeepSeek, and other open-source models.”Available Providers table, Ollama row · checked 2026-09-16

   Visit goose ↗Read the review →Compare gooseGet a quote →From Free
3. #3JU

   Junie

   JetBrains coding agent for the terminal, JetBrains IDEs and CI, running on hosted or your own model keys.

   Best forDevelopers and teams already standardized on a JetBrains IDE who want an agent that shares the same debugger, project index, and version control already built into their editor.

   Open models: Yes — Open-weight models run two ways: through Ollama or LM Studio on any plan, or via the vendor's own Apple Silicon inference engine that ships a quantised Qwen model to your machine.“via OpenAI-compatible APIs, including Ollama and LM Studio”AI models comparison table, sublabel of the 'Local AI models' row, which is ticked in all four plan columns · checked 2026-09-18

   Visit Junie ↗Read the review →Compare JunieGet a quote →From Free

4. Advertisement

   Your agent here. Go Featured to sit above the free listings in this ranking, or buy the slot beside it.

   Get featured · top of your category →
5. #4ZE

   Zed

   Rust-based code editor with a built-in agent panel, ACP external agents and Zed-hosted models.

   Best forDevelopers who want a fast, open-source editor and the freedom to run Claude Code, Codex, Gemini CLI or other vendors' own agents natively via the Agent Client Protocol Zed created, instead of being locked into one editor's built-in agent.

   Open models: Yes — Open-weight models run locally through llama.cpp, Ollama or LM Studio, or any local OpenAI-compatible server; the docs walk through pulling and selecting such a model.“Use local models when you run the model on your machine or on infrastructure you control.”Use a Local Model, opening line and provider table · checked 2026-09-29

   Visit Zed ↗Read the review →Compare ZedGet a quote →From Free
6. #5WA

   Warp

   Agentic development environment: a terminal, a standalone agent CLI, and unattended cloud coding agents.

   Best forDevelopers and teams who want one open platform that runs and coordinates multiple coding agents (Claude Code, Codex, Gemini CLI, Warp Agent) side by side across terminal, CLI and cloud, with shared team context and programmatic API access to cloud agent runs.

   Open models: Yes — Open-weight models are hosted for you through Fireworks AI, and an Auto (Open-weights) router picks between them; Llama is absent from that enumerated table.“Warp also supports leading open source models hosted via Fireworks AI, so you can run them from inside Warp without setting up your own inference infrastructure.”Section "Hosted models (via Fireworks AI)", lead-in sentence · checked 2026-09-03

   Visit Warp ↗Read the review →Compare WarpGet a quote →From Free
7. #6FA

   Factory

   Droid coding agents that run in the desktop app, CLI, IDEs, CI and Factory-managed cloud computers.

   Best forEngineering organizations that want AI agents handling more of the SDLC than code generation alone (triage, testing, release and documentation) and are willing to pay enterprise pricing for deployment flexibility and compliance controls.

   Open models: Yes — Droid Core is an open-weight model pool with its own rate limits, switchable free when standard usage runs out.“When you hit a Standard Usage rate limit, you can switch for free to Droid Core (powered by leading open-weight models) which has a separate rate limit.”FAQ, "What happens if I exceed my limits?" answer · checked 2026-09-02

   Visit Factory ↗Read the review →Compare FactoryGet a quote →From $20/month
8. #7LE

   Letta

   Stateful agent harness, cloud and SDK for agents with persistent, git-versioned memory.

   Best forDevelopers and researchers who specifically need an agent whose memory, skills and identity genuinely persist and evolve across long-running sessions, rather than reset on every call.

   Open models: Yes — Ollama, LM Studio and llama.cpp endpoints are local-only integrations; DeepSeek, Mistral and Together are also listed.“direct integrations with local inference providers such as LM Studio, Ollama, and llama.cpp are available only for local agents”Connecting model providers section, gateway paragraph · checked 2026-09-02

   Visit Letta ↗Read the review →Compare LettaGet a quote →From Free
9. #8CC

   Codex CLI

   OpenAI's terminal coding agent for inspecting, editing and running repository code from the shell.

   Best forDevelopers who already pay for a ChatGPT plan and want a terminal-native agent with granular, inspectable sandbox/permission controls across CLI, desktop and IDE surfaces.

   Open models: Yes — --oss runs Codex against a local Ollama or LM Studio provider.“Codex can run against a local "open source" provider such as Ollama or LM Studio when you pass `--oss`.”OSS mode (local providers) section · checked 2026-09-02

   Visit Codex CLI ↗Read the review →Compare Codex CLIGet a quote →From Free
10. #9OP

    OpenCode

    Open source AI coding agent for the terminal, desktop app and IDE, with 75+ model providers.

    Best forDevelopers who want a genuinely free, fully open-source coding agent that works with any model provider, including their own local Ollama models, without being locked into one vendor's terminal tool.

    Choose OpenCode if you want the widest provider list (75+, including local Ollama) plus a desktop and IDE option alongside the CLI.

    Open models: Yes — Open-weight families such as Qwen, DeepSeek, GLM and Kimi are served through Zen, and local OpenAI-compatible servers can be configured.“OpenCode uses the AI SDK and Models.dev to support 75+ LLM providers and it supports running local models.”Providers page, opening paragraph · checked 2026-09-21

    Visit OpenCode ↗Read the review →Compare OpenCodeGet a quote →From Free
11. #10ZO

    zot

    Single-binary terminal coding agent harness in Go, with 30+ model providers and any-language extensions.

    Best forDevelopers who want a fast, free, Go-native coding-agent harness with real session-branching control and are comfortable running fast-moving, self-described "beta forever" software from an independent maintainer.

    Choose Zot if you want the smallest possible footprint — one static Go binary with 30+ providers built in, no separate install steps.

    Open models: Yes — Open-weight models run locally through Ollama, LM Studio or a llama.cpp router, with in-app GGUF download and load management for the router path and no API key needed.“Managed GGUF models: connect the distinct `llama.cpp` provider to a recent llama.cpp router, then use `/llama` for download, load, unload, and removal.”Providers page, Custom and local providers list · checked 2026-09-28

    Visit zot ↗Read the review →Compare zotGet a quote →From Free
12. #11JU

    JugglerVerified

    Visual workbench for AI coding agents: branching conversation trees, inspectable model transactions, local or remote.

    Best forDevelopers who want hands-on visibility into what an AI coding agent is doing to their codebase and accept early-stage, self-hosted, open-source software.

    Choose Juggler if you want a visual, branching session view over an open-model agent rather than a linear terminal transcript.

    Open models: Yes — Ollama is supported for models running on your own hardware, and DeepSeek, Mistral, Z.AI and OpenRouter cover open-weight models hosted elsewhere; the provider list is explicitly extended by any OpenAI-compatible endpoint.“Juggler supports Anthropic and Claude Code, OpenAI and Codex, GitHub Copilot, Google Gemini, Mistral, Z.AI, Ollama, OpenRouter, DeepSeek and other OpenAI-compatible providers.”README, Model support section · checked 2026-09-17

    Visit Juggler ↗Read the review →Compare JugglerGet a quote →From Free
13. #12OP

    OpenHands

    Open-source software engineering agent with GUI, CLI, SDK and an enterprise control plane.

    Best forOpen source-minded engineering teams and enterprises that want a self-hostable coding agent platform

    Open models: Yes — Open-weight models are supported first-class: a recommended local model plus documented LM Studio, Ollama, vLLM and SGLang serving paths.“We now recommend Qwen3.6-35B-A3B as the first local model to try with OpenHands. It is an open-weight MoE model built for agentic coding”News section, 2026/05/21 entry · checked 2026-09-21

    Visit OpenHands ↗Read the review →Compare OpenHandsGet a quote →From Free
14. #13AI

    Aider

    Open-source terminal AI pair programmer that edits files in your local git repo and commits the changes.

    Best forLowest-friction path to a local or open model

    The lowest-friction path to an open model: point Aider's CLI at any OpenAI-compatible endpoint (Ollama, OpenRouter, a local server) with one config line.

    Open models: Yes — Local open-weight models are explicitly supported through Ollama, LM Studio and any OpenAI-compatible endpoint, so inference can stay on your own hardware.“Aider can work also with local models, for example using Ollama.”Connecting to LLMs, Local models section · checked 2026-09-30

    Visit Aider ↗Read the review →Compare AiderGet a quote →From Free
15. #14CL

    Cline

    Open-source coding agent for VS Code, JetBrains, the terminal and CI, running on whichever model you bring.

    Best forDevelopers and teams who want a transparent, open-source, model-agnostic agent they fully control and pay for by usage, with optional flat-rate access to open-weight models.

    Open models: Yes — Open-weight models are a first-class path: a curated set ships inside the ClinePass subscription, and Ollama, LM Studio or any OpenAI-compatible endpoint runs weights you host.“ClinePass is a low-cost subscription that includes open weight models in Cline.”ClinePass FAQ, 'What is ClinePass?' answer · checked 2026-09-12

    Visit Cline ↗Read the review →Compare ClineGet a quote →From Free
16. #15DE

    Devin

    Cloud AI software engineer that takes a ticket, works in its own VM, and opens a pull request.

    Best forEngineering teams offloading well-scoped migration tasks without adding headcount

    Open models: Yes — Open-weight models including DeepSeek, Kimi and GLM are selectable in Devin CLI.“We also support a number of leading open source models like DeepSeek, Kimi, and GLM.”Available Models section · checked 2026-09-02

    Visit Devin ↗Read the review →Compare DevinGet a quote →From Free
17. #16GC

    GitHub Copilot

    GitHub's AI coding assistant and cloud agent, from editor completions to pull-request review.

    Best forEnterprise and team developers already in GitHub and mainstream IDEs who want governed, reviewable AI coding at scale, with optional access to Claude Code and OpenAI Codex through the same interface.

    Open models: Yes — Two Moonshot AI open-weight models are in the hosted catalog, served by GitHub on Azure AI Foundry and Fireworks AI under zero data retention.“They are open-weight models that may be less aligned than other Copilot models, with an elevated risk of geographic bias.”Open-weight models, Moonshot AI models · checked 2026-09-02

    Visit GitHub Copilot ↗Read the review →Compare GitHub CopilotGet a quote →From Free