[Review] aider pricing: the tool is free, but your LLM bill runs US$10 to US$200+ a month

aider is one of the few coding tools where the free tier question has a genuinely unusual answer: the software itself costs nothing, ever. Apache 2.0 license, no accounts, no trial countdown. But “free” is only half the story. Every time aider edits your code, it calls an AI model through an API — an application programming interface, the paid connection between your terminal and an AI provider’s server. That bill doesn’t go to aider. It goes to whichever provider you plugged in: OpenAI, Anthropic, Google, DeepSeek, or any of 15+ others. The tool is free. The brain is not.

The tool costs nothing; the model does

aider is an open-source, command-line pair programmer. You run it in a terminal, point it at a codebase, and it reads your files, talks to an LLM (large language model — the AI engine doing the actual reasoning), and writes changes directly back to your project. Each edit lands as a Git commit, the standard unit of tracked code change, making every AI action reviewable and revertable. No browser tab required. No cloud account tied to aider itself. Apache 2.0 means commercial use is unrestricted.

Illustration: While the tool chassis is free to pick up, connecting cloud providers routes substantial metered billing from external sources.

What the tool does not bundle: the AI brain. aider supports 15+ LLM providers and works across 150+ programming languages, but every model call carries a cost — unless you route it to a local model running on your own machine.

One provider, very different bills

The spread between a lean setup and an expensive one is wide. Light use — roughly 10 files edited per month — costs about US$10 to US$30 a month depending on your chosen provider and model tier. Heavy use, meaning large codebases or frequent calls to top-tier models like Claude Opus or GPT-4, can reach US$50 to US$200+ a month. aider itself takes no cut; those figures reflect your LLM provider’s API usage rates directly.

There is no aider subscription to stack against those numbers, because aider has no subscription. You’re always paying the underlying model provider, not a product layer on top of it.

The local model escape hatch

For developers who want genuinely zero ongoing cost, local models — most commonly run through Ollama — let aider operate with no external API calls at all. The conversation stays on your machine; nothing leaves for an outside server. The practical limitation is model quality: local models typically trail the cloud frontier in raw coding ability, particularly for complex multi-file edits. But for straightforward refactors, privacy-sensitive codebases, or simply keeping the bill at zero, it is a real option and not just a footnote in the documentation.

🔍 AMPM exclusive check

aider’s official site at aider.chat has no pricing page. There is no URL to a table because there is no table — aider genuinely charges nothing for the software. What the site does not make easy to find is that the cost question gets handed off immediately to your LLM provider the moment you run any command.

The US$10-30 light-use and US$50-200+/month heavy-use figures in our data are cost estimates based on typical API usage patterns, not rates published by aider itself. aider’s documentation references LLM provider costs but does not include a provider comparison table or a cost calculator. A developer who searches “aider pricing” expecting a subscription page will find nothing, and may not immediately realize they need to source and fund an API key separately before the tool can do anything useful.

Nothing we found contradicts the official documentation. But the gap between “this is free” and “here is what you will actually spend” is wider than the site makes apparent, and that gap is where most of the confusion about aider’s real cost lives.

☀️ AMO, the budget-minded cat

aider is one of the most honest free tiers in the coding tool category, because the pricing model is not a stripped-down free plan — there simply is no plan. The software is completely open source, requires no account, and imposes no usage limits of its own. That is genuinely rare in a category where every other tool gates features behind a subscription tier.

The complication is that “free” stops at the terminal prompt. The moment you ask aider to do anything with a cloud model, you’re spending API tokens — the unit LLMs charge by, where one token is roughly a word or word fragment. A light user touching around 10 files a month and using mid-tier models will likely spend US$10 to US$30 a month. That is cheaper than most subscription coding tools, and it scales down gracefully: light months cost less automatically.

The budget discipline required here is different from a subscription, though. A subscription has a hard ceiling; API billing does not. There are no spending alerts built into aider. If you run it on a large codebase with an expensive model and forget to stop, the bill accrues silently in your API account. That is not an aider flaw exactly, but it is a practical risk that flat-rate tool users are not used to managing.

The path to zero ongoing cost is local models. For developers comfortable setting up Ollama, it works and the quality is usable for routine tasks. For everyone else, US$10 to US$30 a month is the realistic floor — which is still worth it for budget-conscious users, provided they treat their API key like a metered utility rather than an unlimited resource.

🌙 PIMI, the performance-minded cat

aider’s model-agnostic design is its strongest technical argument. Supporting 15+ LLM providers means you pick the model that fits the task, not the model a vendor locked you into. For a heavy refactor spanning dozens of files, you route to a top-tier model. For repetitive boilerplate, you drop to a cheaper one. No subscription coding tool gives you that granularity of control over what is actually running your code edits.

The Git-first workflow adds a layer of safety that matters at scale. Every aider edit arrives as a Git commit. That means every change is diff-readable, revertable, and fully visible to any code review tool or audit process that reads Git history. For teams or solo developers who care about traceability, this is a more principled default than tools that write to files with no structured record of what changed and why.

The ceiling on heavy use is a real consideration. aider at full power — top-tier models, large context windows, frequent sessions — can reach US$50 to US$200+ a month in API costs. That figure is not a cap; it can go higher on a particularly intensive sprint. Developers migrating from a flat-rate subscription will find variable billing disorienting at first. Budget forecasting requires tracking API usage separately, outside the tool itself.

The 150+ language support is broad, but practical performance varies by language and by which model you pair with it. aider does not constrain your choices, but the underlying model does, and learning which model works best for your stack takes experimentation. CLI-only operation is also a hard limit: no GUI, no IDE plugin, no web interface. If your workflow depends on an in-editor integration, aider does not replace GitHub Copilot or Cursor.

Verdict: who should use it, who should skip it

Use aider if you’re comfortable in a terminal and want full control over which model handles your code. If you already have an API key with a supported provider, or you run local models, the entry cost is low — US$10 to US$30 a month for light use, or free with Ollama. The Git-native commit structure and support for 150+ languages make it well-suited to developers who treat version control as the ground truth for their work.

Skip aider if you need a GUI or IDE plugin — there is none. Skip it if you want a predictable flat monthly bill; API costs are fully variable and heavy months can reach US$50 to US$200+. Skip it if you have no existing API account and no interest in setting one up, or if you want a vendor support relationship that open-source tools do not provide.

The bottom line: aider is transparent about its cost in a way most tools are not, because it takes no margin at all. What you spend goes directly to the model provider you chose. That is an advantage for developers who want control, and a complexity for anyone expecting a simple subscription.

Written: 2026-10-07 Price last verified: 2026-09-27 Tool last health-checked: 2026-10-07

🔗 Related on AMPM

• aider pricing & free tier
• aider vs bolt
• aider vs claude

📎 Sources

All prices and quotas in this article come from AMPM’s verified dataset for aider (last verified 2026-09-27); primary sources:

• https://aider.chat

🐾 Meet the cats: AMO and PIMI

AMPM is a Taiwan-based AI-tool pricing watchdog. Our motto: ask before you subscribe. Two cats argue every tool from two sides:

• ☀️ AMO — the budget-minded cat. Always asks: is the free tier enough, and is paying actually worth it?
• 🌙 PIMI — the performance-minded cat. Cares about whether it works well, fast, and gets your job done.

When they are done arguing, you know whether to pay. Every number is checked by us, with the verification date at the end.

🐾 Our sister sites also checked

• Travel insurance and flight-delay coverage checked
• Brokerage fees and account opening checked