How to use Continue cheaper without a subscription

Continue is one of the best open-source AI coding assistants for VS Code and JetBrains. The extension itself is free, but the models behind it are not — and the easiest "just works" path nudges you toward a subscription or a per-seat plan. If you mostly want autocomplete, chat, and edits without a recurring bill, there's a cheaper way: bring your own OpenAI-compatible endpoint and pay only for the tokens you actually use.

Why a subscription is often the wrong fit

Subscriptions make sense for heavy daily users with predictable load. But a lot of developers code in bursts: a busy sprint, then a quiet week. A flat monthly fee charges you the same either way, and you're usually locked to one provider's model lineup.

Pay-per-token flips that. You're billed for the requests you make and nothing more. No minimums, no seat counts, no "did I use it enough this month" guilt.

The setup: one endpoint, every model

Continue lets you define any OpenAI-compatible provider in its config. AnyModel gives you a single base URL and one API key that reaches GPT, Claude, Gemini, DeepSeek, GLM, Kimi, Qwen, and Grok. Switching models is just changing the model id — no new keys, no new accounts.

You can start free: 1,000,000 tokens on signup, and a total of 6,000,000 if you link Telegram. No credit card required. After that it's plain pay-per-token, no subscription.

Configure Continue manually

Continue isn't on the one-line installer list (that's for codex, claude, opencode, and hermes), so you set it up by hand. Open your ~/.continue/config.yaml (or config.json on older versions) and add an OpenAI-compatible model block:

models:
  - name: GPT via AnyModel
    provider: openai
    model: gpt-5.1
    apiBase: https://anymodel.org/v1
    apiKey: YOUR_API_KEY
  - name: Claude via AnyModel
    provider: openai
    model: claude-sonnet-4-5
    apiBase: https://anymodel.org/v1
    apiKey: YOUR_API_KEY

The provider: openai line is the trick — Continue talks to AnyModel exactly as it would talk to OpenAI, but every model on the gateway becomes available. Add as many entries as you like; they'll all show up in Continue's model picker.

Quick sanity check

Before wiring it into the editor, confirm the endpoint and key work from your terminal:

curl https://anymodel.org/v1/chat/completions \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"gpt-5.1","messages":[{"role":"user","content":"hi"}]}'

If you get a JSON response back, Continue will too.

Pick the right model for each job

The biggest savings come from matching model to task, not from any single provider. With one key you can mix freely:

Set a small, cheap model as Continue's autocomplete provider and reserve the expensive ones for chat where reasoning matters. Autocomplete fires constantly, so this one change can cut your token spend dramatically. If you're unsure which model wins on price or quality, the compare page lays them out side by side.

A note on privacy

If you're pasting proprietary code into Continue, consider Ghost Mode — opt-in, zero-retention API keys where prompts and responses aren't stored on our side, only a token counter. To be honest about it: the underlying model provider still receives your prompt to generate a reply, so this isn't "100% privacy." It just removes our copy from the equation.

What you actually save

• No monthly fee — you pay for tokens, then stop paying when you stop coding.
• No lock-in — one key, every major model, swap any time.
• A real free tier — up to 6M tokens to test before you spend a cent.

That's enough to run Continue for weeks of light use without touching your wallet. For more setup guides and model breakdowns, browse the blog.

Ready to drop the subscription? Create a free AnyModel account, grab your key, and paste it into Continue's config — you'll be coding in a couple of minutes.