Top 6 of 28 alternatives
Alternatives to Claude Code
Find a better fit. See what each alternative to Claude Code does better, where it falls short, and what it costs.
On this page
- 01
Cursor
Coding agent in a desktop editor, terminal CLI, cloud VMs and CI, priced per seat plus token usage.
From FreeWhy it's on this listTrades Claude Code's terminal for the category's most polished in-editor agent, with model choice across Claude, GPT and Gemini from $20/month.
Where it improves on Claude Code
- +GPT · YesOpenAI models are available in the picker and priced per million tokens from the Other Models pool.Source ↗
- +Open models · PartlyAll inference is hosted: the catalogue includes GLM 5.2 from Z.ai and Kimi K3 from Moonshot, but there is no path to point Cursor at open weights you run.Source ↗
- +Model choice · YesYou pick the model per request, or hand selection to Auto/Cursor Router; each choice bills at that model's own rate.Source ↗
Trade-offs against Claude Code
- −MCP server · PartlyMCP is inbound only: Cursor is the client that connects to MCP servers. Another agent drives Cursor over ACP (agent acp) or the SDKs, not by querying an MCP endpoint Cursor exposes.Source ↗
- −Admin controls · PartlyTeams admins enforce Privacy Mode, team rules and sandbox mode; repository blocklists, model/MCP access controls and auto-run, browser and network controls are Enterprise-only.Source ↗
- −Getting out · PartlyAccount deletion is self-serve and permanent with all data removed within 30 days, and your code stays local; included usage is not portable, resetting each billing cycle.Source ↗
- 02
Warp
Agentic development environment: a terminal, a standalone agent CLI, and unattended cloud coding agents.
From FreeWhy it's on this listGPT · YesOpenAI models are listed in the vendor's own model table alongside Anthropic and Google; Codex is also supported as an alternative harness.Source ↗
Where it improves on Claude Code
- +Open models · YesOpen-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.Source ↗
- +Model choice · YesYou pick from the model picker or write your own routing rules; an admin or your plan can grey out models, and Auto hands selection back to Warp if you want it.Source ↗
Trade-offs against Claude Code
- −In your editor · NoThere is no first-party IDE or editor extension; Warp ships its own editor and diff review and positions itself as the thing that replaces the external IDE. It can only launch your editor.Source ↗
- −MCP server · PartlyWarp is a mature MCP client for local and cloud agents. Being queried by another agent works only through the hosted Factory MCP server, which is inside Warp Factories Early Access.Source ↗
- −Admin controls · PartlyAdmins set org-wide autonomy per action type plus regex command allowlists and denylists, and can lock users out of overriding them. On Free most settings are fixed; configurability starts at Business and is complete on Enterprise.Source ↗
- 03
Augment Code
Cosmos platform running fleets of coding agents across the SDLC, plus Auggie CLI and IDE extensions.
From $100/monthWhy it's on this listGPT · YesGPT-5.1 through GPT-5.6 variants are selectable, and a GPT Prism router is offered.Source ↗
Where it improves on Claude Code
Trade-offs against Claude Code
- −API access · PartlyAuggie TypeScript/Python SDKs drive the agent from your own code; the Analytics API is Enterprise-only preview.Source ↗
- −Bring your own key · PartlyYou can supply your own machines or cloud for execution, but all inference is billed through Augment; no bring-your-own model key is documented.Source ↗
- −Opt out of training · PartlyPaid plans exclude AI training on your data, but the Community terms grant Augment the right to train on Customer Code.Source ↗
- 04
Devin
Cloud AI software engineer that takes a ticket, works in its own VM, and opens a pull request.
From FreeWhy it's on this listGoes further than Claude Code's interactive sessions into fully async delegation — assign a ticket, walk away, review the pull request it opens.
Where it improves on Claude Code
- +GPT · YesOpenAI models are selectable in Devin CLI, with gpt among the three the vendor recommends trying first.Source ↗
- +Open models · YesOpen-weight models including DeepSeek, Kimi and GLM are selectable in Devin CLI.Source ↗
- +Open source · PartlyOnly the Outposts Kubernetes orchestration operator is published as open source; no source distribution is documented for Devin itself.Source ↗
Trade-offs against Claude Code
- −Bring your own key · PartlyYour own AWS KMS key encrypts data at rest, but only on the Enterprise Assured deployment; no bring-your-own-inference is documented.Source ↗
- −Opt out of training · PartlyTraining is on by default and the opt-out is paid-plans-only; opting out also enables Zero Data Retention with the model providers.Source ↗
- −Getting out · PartlyPurchased on-demand credits roll over and never expire; no account-wide data export is documented beyond invoices and per-resource API reads.Source ↗
- 05
Junie
JetBrains coding agent for the terminal, JetBrains IDEs and CI, running on hosted or your own model keys.
From FreeWhy it's on this listGPT · YesOpenAI models are reachable both through the JetBrains subscription's curated cloud set and by connecting your own OpenAI key directly, with built-in CLI aliases routing to them.Source ↗
Where it improves on Claude Code
- +Open models · YesOpen-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.Source ↗
- +Model choice · YesYou choose, not the vendor: four provider classes are supported and a flag forces a specific one, though leaving the default lets JetBrains pick on price-quality grounds.Source ↗
Trade-offs against Claude Code
- −In a browser · PartlyThe web app is a companion view onto a CLI session still running on your own machine, not a standalone hosted agent, and the vendor lists a dozen features it cannot do.Source ↗
- −MCP server · PartlyMCP runs one way only: Junie consumes external MCP servers as a client. Another program reaches it through ACP instead, and no MCP server exposing Junie is documented.Source ↗
- −Admin controls · PartlyA developer can constrain the agent locally through the allowlist file on any tier, but organisation-wide policy such as company MCP rules is an enterprise capability.Source ↗
- 06
Codex CLI
OpenAI's terminal coding agent for inspecting, editing and running repository code from the shell.
From FreeWhy it's on this listOpenAI's own terminal-native agent — bundled into any ChatGPT plan you may already pay for, even a limited Free tier, instead of a separate subscription decision like Claude Code's.
Where it improves on Claude Code
- +GPT · YesGPT-5.6 Sol, Terra and Luna are all marked available in Codex CLI, selected with --model.Source ↗
- +Open models · Yes--oss runs Codex against a local Ollama or LM Studio provider.Source ↗
- +Open source · YesCLI, SDK and app server are in openai/codex; the IDE extension and Codex cloud are not open source.Source ↗
Trade-offs against Claude Code
- −Whole-codebase work · PartlyDocumented for repository-wide exploration and multi-step plans; the vendor states no repo-wide migration guarantee.Source ↗
- −Claude · NoSign-in and model support are OpenAI models only; no Anthropic model appears in any documented model list.Source ↗
- −MCP server · PartlyCodex is an MCP client for local and remote servers, but exposing Codex itself as an MCP server is deprecated in favour of the app server.Source ↗
Side by side
Compare Claude Code with the shortlist
Compare alternatives to Claude Code
Pick up to three alternatives to see how they differ from Claude Code.
The digest
Which agents actually ship. New agents and articles, by email, at most once a week.
Advertisement
Reach buyers choosing an agent. Billed monthly. Cancel anytime.
See ad options →Featured · top of your category →The digestFree
Which agents actually ship.
New agents and articles, by email, at most once a week.
One-click unsubscribe