Cline’s pricing page still has no subscription on it: the extension is free, and what you pay by default is whatever your model provider charges for the tokens Cline sends it. The one subscription Cline does sell, ClinePass, lives on its own page and is optional.
That structure comes from how Cline is built. It’s Apache 2.0 open source, distributed as a VS Code extension and, on Enterprise, a JetBrains plugin. Install it and you get the full agent: file edits, terminal command execution, an MCP marketplace for extending its tool access, multi-root workspace support, all without a paywall. The catch, and it’s not really a catch, is that agentic coding runs on LLM inference, and inference costs money regardless of which tool is driving it.
You have three ways to pay for the inference itself. Bring your own API key from Anthropic, OpenAI, Google, OpenRouter, AWS Bedrock, GCP Vertex, Groq, Cerebras, DeepSeek, or a local model through Ollama or LM Studio, and you’re billed directly by that provider at their standard rates. Use Cline’s own provider option, which sells inference “at cost” without adding a margin on top. Or, since June 2026, subscribe to ClinePass: $9.99 a month for a curated set of open-weight models (GLM, Kimi, DeepSeek, MiniMax, MiMo and Qwen) at two to five times the standard API rate limits, metered against 5-hour, weekly and monthly windows. Cline’s docs are explicit that it’s optional, and the page adds that “additional processing fees may apply.” With the first two routes, your monthly cost scales exactly with how much you use the agent and which models you route work through, not with a fixed tier you might over- or under-use.
For solo developers and small teams, this is close to the best-case pricing structure if you’re already comfortable managing API keys and watching a token bill. Heavy users running large agentic sessions against a frontier model like Claude Opus will spend real money, potentially more per month than a flat $20 subscription elsewhere, but light users pay close to nothing beyond occasional API calls. ClinePass is the predictable-bill option for people happy to stay on open-weight models; it doesn’t cover frontier closed models.
Enterprise exists for organizations that need more than “install the extension and expense the API bill.” It adds a JetBrains extension alongside VS Code, SSO, centralized billing so finance doesn’t have to reconcile a dozen individual API accounts, role-based access control, the ability to limit which inference providers employees can route through (useful for data governance), a team management dashboard, and authentication logs. None of that changes how inference itself is billed; it changes who can see and control the spending. Pricing here is a custom quote through Cline’s sales team, gated behind a “Contact Sales” flow rather than a published number, which is standard for this tier of security and admin tooling across the category.
The practical decision is less about which Cline tier to pick and more about which model provider and spending controls fit your situation. If you’re an individual developer or a small team that trusts everyone with their own API key, Open Source covers you completely and costs nothing beyond usage. If you’re rolling this out across an engineering org and need to control which providers people can hit, track spend centrally, or meet an SLA, that’s what pushes you into an Enterprise conversation. Enterprise is an administrative and compliance add-on, not a feature unlock.
One thing worth flagging for anyone comparing Cline against subscription-priced competitors: outside ClinePass, Cline’s site publishes no usage-based pricing table beyond “pay at cost or BYOK,” so your actual monthly cost depends on your model choice and usage pattern, not on anything Cline controls. Budgeting for Cline on those routes means budgeting for your model provider’s API pricing, plus zero markup if you use Cline’s own inference option.