AI · Guide

AI Coding Tools Free to Try: Five of Six, Not Claude Code

GitHub Copilot Free gives 2,000 completions a month; Antigravity replaced Google’s free Code Assist. Paid plans start at $8.

Nadia RoyAI analyst3 min · 716 words · no filler
Free tiers of six AI coding tools, from each maker’s pricing page, read 16 October 2026, with US prices; free allowance, then cheapest paid plan. GitHub Copilot: 2,000 completions a month; Pro, $10 a month. Google Antigravity: unlimited Tab, weekly rate limits; Google AI Pro, $19.99 a month. Cursor: limited agent and chat requests; Plus, $8, Grok only. Devin Desktop: light agent quota, unlimited Tab; Pro, $20 a month. OpenAI Codex: limited, on ChatGPT Free; Go, $8 a month. Claude Code: not on Claude Free; Pro, $20 a month. No free tier here states an agent request count.
Illustration: bitcritiq
Reading mode

If you buy through our links, we may earn a commission. It never affects our verdicts or scores — how that works. As an Amazon Associate I earn from qualifying purchases.

Five of the six AI coding tools we checked have a free tier: GitHub Copilot, Google’s Antigravity, Cursor, Devin Desktop, where Windsurf’s pricing page now leads, and OpenAI’s Codex through ChatGPT Free. Claude Code is the exception; Anthropic includes it from the $20 Pro plan up. GitHub is the only maker of the six that puts a number on a free allowance: 2,000 code completions a month.

The others describe their free caps in words: “limited”, “light”, “basic”. None of the six publishes a free count for chat or agent requests, the part that writes and edits code across files rather than finishing the line you are typing. Every limit and price below is from the maker’s own page, read on 16 October 2026, in US dollars.

What each free tier includes

Tool Free allowance Paid entry price Notes
GitHub Copilot Free 2,000 completions a month; limited chat and agent use Pro, $10 a month Haiku 4.5, GPT-5 mini “and more”; no credit card
Google Antigravity Unlimited Tab completions and command requests; basic weekly rate limits Google AI Pro, $19.99 a month Replaced Gemini Code Assist for individuals
Cursor Free Limited agent and chat requests Plus, $8 a month Plus is Grok models only
Devin Desktop, Free Light agent quota; unlimited inline edits and Tab completions Pro, $20 a month Limited model availability on Free
OpenAI Codex Limited access on ChatGPT Free ChatGPT Go, $8 a month Go is also listed as limited; Plus at $20 expands it
Claude Code Not included on Claude Free Pro, $20 a month $17 a month billed yearly

The two that changed this year

Google’s free coding tier moved. Google says Gemini Code Assist stopped serving requests for its free individuals tier on 18 June 2026, and the same applies to Gemini CLI; the Login with Google option for both is gone. Google points those users to Antigravity, whose $0 plan lists unlimited Tab completions and command requests under “basic weekly rate limits”. The next step up is Google AI Pro, $19.99 a month on Google’s US subscriptions page.

Windsurf’s pricing page now opens Devin’s. The $0 plan there is for Devin Desktop and lists “a light quota to code with agents”, limited model availability, and unlimited inline edits and Tab completions. Pro at $20 adds frontier models from OpenAI, Anthropic, Google and SpaceXAI.

Where the limits bite

Copilot. Free covers editors and IDEs, the Copilot CLI, github.com and agent mode in VS Code, Visual Studio, JetBrains, Eclipse and Xcode. It leaves out the cloud agent that turns an issue into a pull request, and pull request code review; both start at Pro. Chat and agent use draw on GitHub AI Credits, which Free gets in a limited amount and Pro gets as $15 a month.

Cursor. Its pricing page now sells Cursor inside X subscriptions, with one usage pool shared across Cursor and Grok that refreshes weekly. The $8 Plus plan is the cheapest way past Free, but it includes Grok models only. From Premium at $30 a month, models from other makers are available at API list price, billed on top.

Codex and Claude Code. OpenAI lists “Limited Codex access” on ChatGPT Free without a number. Anthropic’s comparison table marks Claude Code “No” on Free. Our guide to what Claude Pro and Max add covers the paid side, and the free chatbot tiers set out what the same makers give away for plain chat.

The free option that runs on your machine

A local model has no subscription and no weekly cap. Running one takes a free app such as Ollama or LM Studio, but for coding work the limit becomes your hardware: context length is where local coding hits the wall, because Ollama’s own documentation asks for far more context for coding agents than it gives a smaller graphics card by default.

None of this says which tool writes better code. It says what each maker gives away before you pay, and how little of it the makers count in numbers.

Questions people ask

Is GitHub Copilot free?

Yes, with limits. Copilot Free costs $0 with no credit card and includes 2,000 code completions a month, limited chat and agent use, and models including Haiku 4.5 and GPT-5 mini. Pro, at $10 a month, makes completions unlimited.

Can I use Claude Code for free?

Not on Anthropic’s own plans. Claude’s pricing page marks Claude Code “No” for Free and “Yes” for Pro, which is $20 a month billed monthly or $17 a month billed yearly.

Is OpenAI Codex free?

In a limited way. OpenAI’s pricing page lists “Limited Codex access” on ChatGPT Free and on Go at $8 a month. Plus, at $20 a month, has expanded Codex usage. OpenAI does not publish a free task count.

What happened to Gemini Code Assist for individuals?

Google stopped serving it on 18 June 2026, along with Gemini CLI sign-ins on personal accounts. Google points those users to Antigravity, whose $0 plan lists unlimited Tab completions and basic weekly rate limits.

Is Windsurf still free?

windsurf.com/pricing now opens Devin’s pricing page, which lists a $0 plan for Devin Desktop with a light quota for agents, limited model availability, and unlimited inline edits and Tab completions. Pro is $20 a month.

Which free AI coding tool gives the most usage?

The makers do not publish the numbers needed to compare them. GitHub is the only one that states a figure, 2,000 completions a month, and Antigravity and Devin Desktop list unlimited Tab completions. None of the six states a free agent or chat request count.

How we researched this

No one at bitcritiq has handled this product. Everything here comes from published sources, listed below.

What this cannot tell you
We did not sign in to any of these tools, so this is what the makers publish, not what an account shows. Only GitHub gives a number for any free allowance, and that number covers code completions, not chat or agent use; the other caps are described in words such as ‘limited’, ‘light’ and ‘basic’, so the tiers cannot be compared on size. Cursor’s pricing page now sells Cursor inside X subscriptions alongside Grok, and we report the plans as listed. Devin’s pricing page does not use the name Windsurf; the link from windsurf.com is our only evidence of the connection. Antigravity’s pricing page names Google AI Pro and Ultra as its paid plans and does not mention the cheaper Google AI Plus, so we list Pro as the step up. We have not used or compared the quality of any of these tools.
How we chose this, and what we did
Why this subject
Searches for AI coding tools rose about 70% and for Claude Code with a local LLM about 130% over the six Google Trends pulls we ran, and ‘best local llm for coding 2026’ was a Breakout. The question underneath is what each tool gives you before you pay. Our existing posts cover free chatbots and Claude’s plans, not coding tools.
How we looked at it
Every figure is from the makers’ own pages, read in a browser on 16 October 2026: GitHub’s Copilot plans page, including its comparison table; Antigravity’s pricing page; Google’s developer notice on Gemini Code Assist consumer accounts; Google’s US AI subscriptions page; Cursor’s pricing page and its FAQ text; Devin’s pricing page, where windsurf.com/pricing now lands, and its FAQ; claude.com/pricing, including its comparison table; and chatgpt.com/pricing, including its comparison table. For Claude Code’s absence from Free, the control is that the same comparison table marks it Yes for Pro, Max 5x and Max 20x.

What this rests on

6 claims, all official. Nothing here was measured by bitcritiq — see how we test for why. Open a claim to read the source it came from.

  • GitHub Copilot Free includes 2,000 completions a month and limited chat and agent use, with Haiku 4.5 and GPT-5 mini among its models; Pro costs $10 a month.Official

    GitHub Copilot plans page: Free “$0 USD … 2,000 completions per month · Access to Haiku 4.5, GPT-5 mini, and more · Copilot CLI”; comparison table, Total GitHub AI Credits: “Limited chat and agent usage”; Pro “$10 USD per user / month”.

  • Gemini Code Assist for individuals stopped serving requests on 18 June 2026, including through Gemini CLI, and Google points those users to Antigravity.Official

    Google for Developers: “Starting June 18, 2026, Gemini Code Assist IDE extensions stopped serving requests for the Gemini Code Assist for individuals, Google AI Pro, and Google AI Ultra tiers. This also applies to usage of Gemini CLI … can migrate to the Antigravity family of products.”

  • Antigravity’s $0 plan includes unlimited Tab completions and command requests, with basic weekly rate limits.Official

    Antigravity pricing: “For Individuals $0/month … Access to a variety of agent models · Unlimited Tab completions · Unlimited Command requests · Basic weekly rate limits”.

  • Cursor’s free plan has limited agent and chat requests; its $8 Plus plan includes Grok models only, and third-party models are billed at API list price from Premium up.Official

    Cursor pricing: Free “Limited agent requests · Limited Chat requests”; FAQ: “Plus includes Grok models only. Premium and above include generous usage of Grok models but do not include an allowance for third-party models. Third-party models remain available in Cursor at API list price.”

  • Claude Code is not included on Claude’s Free plan; Pro, which includes it, costs $20 a month billed monthly.Official

    claude.com/pricing comparison table: “Claude Code — No (Free), Yes (Pro), Yes (Max 5x), Yes (Max 20x)”; Pro: “$17 Per month with annual subscription discount ($200 billed up front). $20 if billed monthly.”

  • ChatGPT Free includes limited Codex access; Plus at $20 a month expands Codex usage.Official

    chatgpt.com/pricing: Free “Limited Codex access”; Plus “$20 / month … Expanded Codex usage”; comparison table Codex row: “Limited” for Free and Go, “Expanded” for Plus.

Sources 8

  1. GitHub Copilot · Plans & pricing — GitHubOfficialaccessed Oct 16, 2026
  2. Google Antigravity — PricingOfficialaccessed Oct 16, 2026
  3. Gemini Code Assist consumer accounts — Google for DevelopersOfficialaccessed Oct 16, 2026
  4. Google AI subscriptions (United States) — GoogleOfficialaccessed Oct 16, 2026
Show 4 more · 4 official or standards
  1. Cursor · Pricing — AnysphereOfficialaccessed Oct 16, 2026
  2. Plans and Pricing — Devin (reached from windsurf.com/pricing)Officialaccessed Oct 16, 2026
  3. Plans & pricing — Claude by AnthropicOfficialaccessed Oct 16, 2026
  4. Plans and Pricing — ChatGPTOfficialaccessed Oct 16, 2026

Pass it on

Know someone weighing this up? Send them the answer.

bitcritiq.com/69xk

The link is permanent: it opens this article for good, even if the title changes.

read next

Hermes Agent: what it needs to run. Nous Research’s MIT-licensed AI agent, from its README and docs, read 16 October 2026. With an API: a cloud provider’s key, or one Nous Portal sign-in. With a local model: Ollama, vLLM, llama.cpp or LM Studio, with at least 64,000 tokens of context. By default: commands run on your machine with no isolation; the docs advise Docker for production. Talk to it from a terminal, Telegram, Discord, Slack and more.
AI · Explainer

Hermes Agent Is Nous Research’s Open-Source AI Agent

It runs on your machine, a server or a cloud sandbox, and uses a cloud API or a local model with at least 64,000 tokens of context.

OpenClaw can run on a local model, from OpenClaw's and Ollama's documentation, read 14 October 2026. Ollama: use the native address; the /v1 address breaks tool calling. LM Studio: its local server at 127.0.0.1:1234, Responses API. llama.cpp: a server OpenClaw manages, chosen in onboarding. Needs: tool calls that work and 64k or more context, per Ollama. Local models skip hosted safety filters, so tighten tool permissions.
AI · Explainer

OpenClaw Runs on a Local LLM via Ollama or LM Studio

OpenClaw's docs also cover llama.cpp, vLLM and any OpenAI-compatible server. Ollama advises 64k context; small models are easier to hijack.

What the free plans cap, and what paid costs. US list prices from each maker’s own page, read 12 October 2026. Grammarly Free: 100 AI prompts a month; paid, $30 a month or $144 a year. QuillBot Free: paraphrases of up to 125 words; paid, $19.95 a month or $99.95 a year. Wordtune Basic: 10 rewrites and AI suggestions a day; paid, $13.99 a month, or $6.99 a month billed yearly. ProWritingAid Free: 10 Rephrases a day, and reports on up to 500 words; paid, $30 a month or $120 a year. Microsoft Editor: basic spelling and grammar; paid, $9.99 a month or $99.99 a year. Cheapest paid plan from each, before tax; Microsoft’s is Microsoft 365 Personal. Not a ranking.
AI · Explainer

Free AI Writing Tools Ration Rewrites in Different Units

Grammarly gives 100 AI prompts a month, Wordtune 10 rewrites a day, QuillBot 125 words a paraphrase. Yearly billing halves most prices.

Specifications