Aider pricing: a free client and the API bill you actually pay

Aider, the open-source AI programming assistant for the terminal, costs $0. It is Apache-2.0 software, and aider.chat has no plans, tiers, seats, or hosted service, so “Aider cost” effectively means the model API charges incurred through it. What you should estimate is tokens per session multiplied by the provider's rate. The setting most likely to inflate that number is Aider's default behavior when it encounters a model ID it does not recognize—two lines fix it, as explained below.

First, a distinction, because search results mix the two up: another company that makes accounting close software for firms is also called Aider and publishes a monthly minimum price. Those prices belong to that product and have nothing to do with this CLI. This page covers only the CLI at aider.chat, the tool distributed as the aider-chat package from the Aider-AI/aider repository.

Who you pay, and what you pay for

Aider's FAQ says the tool is released as Apache 2.0 open-source software. The homepage, FAQ, installation page, and LLM connection page contain no pricing, plans, or subscription language (verified October 1, 2026). All billing happens on the other side of the endpoint.

Before standardizing on it for a team, check the update status. The repository is not archived and has 49,309 stars, but the last commit on main was May 22, 2026 (GitHub API, October 1, 2026). The two official “latest version” listings also differ: PyPI lists 0.86.2, uploaded February 12, 2026, while GitHub Releases stops at v0.86.0, dated August 9, 2025. pip install installs the PyPI version; use that version as the reference. For Python versions, use the >=3.10,<3.13 constraint in the 0.86.2 package metadata as the reference, since that is the condition pip actually checks.

The setting that most affects API costs

This is the part a price list won't show. Aider matches the model ID against a built-in table to determine edit format, repository map (repo map), and caching behavior. It matches the entire string after --model, including the prefix. According to the advanced model settings page, the defaults when nothing matches are edit_format: whole, use_repo_map: false, cache_control: false, and weak_model_name: null.

In the source for release 0.86.2, matching happens in two steps. First it checks for an exact match among 313 full names, such as gemini/gemini-3-pro-preview and openai/gpt-5.2. Only if none match does it try substring rules, including sonnet-4-5, haiku-4-5, and gpt-4.1. For an OpenAI-compatible gateway, the model ID must be passed as openai/<id>. The 40 entries in the table with a openai/ prefix are all OpenAI's own model names. So openai/claude-sonnet-4-6 matches neither step, and runs with whole-file rewrites and the repository map disabled—the model must output each edited file in full on every turn, and output tokens are the more expensive side of the bill.

Two lines fix it. Aider reads this file from the home directory, Git root, or current directory. The key is the full name passed to --model:

# name 必須和傳給 --model 的字串完全一致（連 openai/ 前綴一起）
- name: openai/claude-sonnet-4-6
  edit_format: diff
  use_repo_map: true

For a one-time run, you can also add --edit-format diff at launch. Conversely, openai/claude-haiku-4-5 matches the substring haiku-4-5 and gets diff format and the repository map automatically. This is a coincidence in the name and could change with any Aider version, so verify it yourself rather than relying on it.

One more easily overlooked effect: Aider calls the API in three roles—the main model, an editor model for --architect mode, and a weak model for commit messages and conversation-history summaries (see the options reference). When the ID is unrecognized, the source sets both weak and editor to the main model, so all three roles are billed at the most expensive rate. Use --weak-model to specify a cheaper ID and avoid this (the openai/ prefix is required here too).

Finally, an unrecognized ID also hides costs. Without input_cost_per_token, Aider prints only the token count and no cost line, and the warning page says it “assumes the model is free.” Restore the display with the following metadata file:

{
  "openai/claude-sonnet-4-6": {
    "input_cost_per_token": 0.0000021000,
    "output_cost_per_token": 0.0000105000,
    "litellm_provider": "openai",
    "mode": "chat"
  }
}

Even with metadata, the amount Aider shows is its own calculation based on reported tokens, and the cache multiplier is hard-coded for a specific provider. Treat it as a reference and confirm against the provider's usage records.

Estimated cost for one session

The following is token arithmetic, not a measured task cost or a billing cap. It assumes one session sends 200,000 uncached input tokens and receives 12,000 output tokens using diff edits. The whole-file column assumes the same task produces 40,000 output tokens because of full-file rewrites (an illustrative ratio, not a measured value). Rates are current prices per million tokens from the Kunavo catalog.

Under this assumption, the editing format alone changes the cost of the same Claude Sonnet 4.6 task from $0.546 to $0.840, so correct the configuration before comparing provider rates. If commit messages and summary traffic (assuming 30,000 input and 2,000 output tokens) remain on the main model, Claude Sonnet 4.6 costs $0.084; assigning the weak model to Claude Haiku 4.5 brings the cost to $0.028.

Multiply by your number of sessions per day to turn this into a budget. Kunavo catalog amounts are a billing floor, not a ceiling. When the upstream provider reports costs, you are charged the higher of the catalog amount or upstream cost × the applicable margin. Cache charges and external tools are not included in this example.

“Cheapest API” is two different questions

Aider has one verifiable benchmark for task costs: its polyglot benchmark has models solve the same 225 Exercism problems in 6 languages under the same conditions and records the cost.

The figures come from the leaderboard data file. Two caveats matter more than the ranking: the latest entry is October 3, 2025, so later models have never been measured here. The cost column gives no price list or date, so it shows the cost of that run at the time, not a rate available today. These models are cited as benchmark evidence; DeepSeek is not in the Kunavo catalog. The lower-priced model in the estimate table above is Claude Haiku 4.5.

The practical takeaway is this: o3-pro cost five times as much as gpt-5 (high) and scored lower, so a higher unit price does not guarantee completion. A lower unit price does not guarantee the lowest total either: a cheap model that needs three retries can cost more than an expensive one that finishes in a single attempt. Choose the least expensive model that completes the task with an acceptable review burden, then verify it against the actual cost recorded in your account for that task. Aider's /tokens is, according to the command reference, the token count in the current conversation context, not the total amount billed so far.

Who each route is for

Two limits of the gateway route. Gateways without an Aider-specific module connect using OPENAI_API_BASE, OPENAI_API_KEY, and the required openai/ model ID prefix, as described on the OpenAI-compatible API page. The documentation does not describe a setting to override the Anthropic-format base URL, and the string ANTHROPIC_API_BASE does not appear in the 0.86.2 source. Do not expect prompt-cache discounts on this route: the caching page lists specific supported providers, requires --cache-prompts to be explicitly enabled, and says cache statistics and costs are unavailable during streaming. This page has not tested whether caching works on the compatible route.

Using Aider with Kunavo, and how to pay in Taiwan

Kunavo has a setup guide for this client, but it is a public configuration reference, not a compatibility test. We have not actually run Aider against the Kunavo endpoint; all the information on this page comes from Aider's official release source and documentation. Keep your current working route, run one well-scoped task, and check the actual amount recorded in your account.

Payment is prepaid top-up with no monthly fee. The minimum top-up is $10; this is the minimum amount to activate the balance, not a task fee or subscription. Checkout is handled by Stripe. In Taiwan, you can use credit cards (Visa, Mastercard, American Express, JCB, UnionPay), Apple Pay, Google Pay, and Link; local Taiwanese payment methods such as JKOPay and LINE Pay are not currently available. See the billing information for details, and when ready, you can create an account. For the fuller English original, see Aider pricing.

Frequently asked questions

Aider's repository, package, documentation, and leaderboard data were verified on October 1, 2026. Behavior descriptions are based on the source for the official v0.86.2 release. Kunavo rates are taken from the live catalog, and all price examples on this page are illustrative token calculations, not measured task costs.