You want suggestions while you type
Copilot Pro. GitHub documents unlimited inline and next edit suggestions on Pro that do not use AI credits. For Claude Code, completions as you type are not documented on any plan, so that is not verified.
AI Compare
Prices for United States, before sales tax Data release 2026.10.16-3, Oct 16, 2026 Rules 2.1.0
Comparing Claude Pro and Copilot Pro: individual plans, prices for United States. Change the plans below to compare any other two.
This page is for developers choosing between Claude Pro from Anthropic and GitHub Copilot Pro from GitHub. We compare these two because each is the lowest plan of its product that documents an agent working in the cloud and pull request review. Claude Pro, at $20 a month or $200 a year paid up front, is the first Claude plan with Claude Code, its cloud sessions and reviews through GitHub Actions. Copilot Pro, at $10 a month, is the first Copilot plan with the cloud agent and code review. We also cover Claude Free and Max, and Copilot Free, Pro+ and Max where they change the choice.
Everything here comes from Anthropic's and GitHub's own pricing pages and documentation, checked on October 12 and 13, 2026. We have not tested either tool yet, so nothing on this page says which one writes working code. Where a vendor does not document something, we say it is not verified, which is not the same as documented as not offered.
$20 a month or $200 a year
$10 a month
Opus, Sonnet, Haiku; Fable via usage credits.
'A selection of models': Claude Sonnet 5.5, Haiku 5.5, GPT-5.6 Terra/Luna, GPT-6 Luna, GPT-5.4, Gemini 3.8 Flash, Grok 4.7, Kimi K3, MAI-Code. Not included: Claude Opus/Fable, GPT-5.6 Sol, GPT-6/6.1 Sol.
Up to 1M on all individual plans.
Optional 1 million token context for supported models, in VS Code and Copilot CLI only; uses more AI credits. Default context size per model not stated on the pages checked.
Pricing comparison lists only Anthropic models for Pro: Fable (uses 50% of weekly limits), Opus, Sonnet, Haiku. No models from other AI companies are offered.
Choose models from Anthropic, OpenAI, Google, xAI, Moonshot (Kimi) and Microsoft (plan decides which); auto selection also offered with a 10% model-cost discount.
Claude chat on web, desktop and mobile for writing, editing, questions and everyday work (at least 5x Free usage per 5-hour session); chat and Claude Code share one usage pool.
Docs scope Copilot to software work: code, repositories and development concepts. General questions shown are programming-related; not positioned as an everyday writing assistant.
Web search on all plans.
Optional Bing web search for Copilot Chat (off by default; enable 'Copilot access to Bing' in settings). Setting listed for Free, Pro, Pro+ and Max. Display of source links not confirmed on the page fetched.
Research.
No research-report mode documented; cloud agent can do 'deep research of a repository' (code research only, paid plans).
PDF, DOCX, CSV, TXT, XLSX etc.; 500MB per file, up to 20 files per chat (no plan split stated).
Attach images and PDFs to chat prompts (GitHub.com, IDE, CLI) when using a model that supports image input.
Create and edit files with code execution.
Spreadsheet/CSV analysis with computation not documented as a feature.
Claude Code included; 5-hour session + weekly limits.
Agent mode in VS Code, Visual Studio, JetBrains, Eclipse and Xcode, plus Copilot CLI; consumes AI credits.
Claude Code extension for VS Code and Cursor (also other VS Code forks such as Devin Desktop, Kiro) and plugin for JetBrains IDEs (IntelliJ IDEA, PyCharm, Android Studio, WebStorm, PhpStorm, GoLand). Shares Pro usage limits.
Extensions for VS Code, Visual Studio, JetBrains IDEs, Xcode, Eclipse, Neovim, Zed, SQL Server Management Studio and Raycast; features vary by IDE (see feature matrix).
Claude Code is an agent (chat panel, inline diffs) in VS Code/JetBrains; no page documents inline as-you-type code completions. The Desktop app docs explicitly say it has no autocomplete; the IDE extension docs are silent.
Inline suggestions and next edit suggestions are unlimited and do not consume AI credits.
Claude Code on the web: cloud sessions run on Anthropic-managed VMs, keep running after you close your laptop, clone GitHub repos and create PRs; auto-fix PRs. Counts toward Pro usage limits; no separate VM charge.
Copilot cloud agent researches a repo, plans and codes in a cloud dev environment and opens a PR; GitHub-hosted repos only, one repo/branch/PR per task, 59-minute session limit; uses AI credits.
Via Claude Code GitHub Actions authenticated with a Pro subscription token: review workflow posts inline PR comments automatically. Managed Code Review (research preview) is Team/Enterprise only; local /code-review also works.
Copilot code review on GitHub PRs (manual or automatic for PRs you create) and file-diff review in editors; ~$0.05-$1 (Lite) or $0.25-$5 (Balanced) of AI credits per review.
No photo/illustration generation; can build diagrams/charts/SVG visuals (beta, web/desktop).
Image generation not documented for Copilot.
Claude Motion (beta) marked No on Free, Pro, Max.
Video generation not documented for Copilot.
Voice mode on all plans.
Only voice input (dictating prompts) is documented for Copilot CLI; no spoken two-way conversation documented.
Anthropic says Claude does not generate photos or illustrations; it can view and analyze uploaded images. No page documents editing an uploaded photo.
No GitHub page documents image generation or editing in Copilot; the plans page lists image input only.
No Anthropic page documents background removal. Claude does not generate photos or illustrations.
No GitHub page documents background removal in Copilot.
No image generation: Claude does not make photos or illustrations, so there are no generated images to put text in. It builds HTML and SVG diagrams, charts and visuals instead.
No GitHub page documents image generation in Copilot, so text in images is not described.
No logo maker or logo template is documented. Claude Design (all plans) makes visuals, mockups, prototypes, one-pagers and landing pages.
No GitHub page documents a logo maker in Copilot.
Custom visuals (beta, web and desktop, all users) are built with HTML and SVG; you can download a visual as an .svg or .html file.
No GitHub page documents vector or SVG export as a Copilot feature.
No Anthropic page documents image upscaling for Claude.
No GitHub page documents image upscaling in Copilot.
Claude Motion (short animations built as code from your text, charts and images, exported as MP4) is marked No on Free, Pro and Max; it is a Team and Enterprise beta and uses no video generation model.
GitHub Copilot is a coding assistant; its plans page lists no video features and no bundled video app.
No Claude page documents editing uploaded video footage on Free, Pro or Max; Claude Motion edits only its own code-built animations and is not on these plans.
GitHub Copilot is a coding assistant; its plans page lists no video features and no bundled video app.
No Claude page documents a digital presenter. Claude Motion (not on this plan) 'doesn't create realistic footage or people'.
GitHub Copilot is a coding assistant; its plans page lists no video features and no bundled video app.
No Claude page documents lip sync on any plan.
GitHub Copilot is a coding assistant; its plans page lists no video features and no bundled video app.
No Claude page documents automatic video captions on any plan.
GitHub Copilot is a coding assistant; its plans page lists no video features and no bundled video app.
No Claude page documents turning long videos into short clips on any plan.
GitHub Copilot is a coding assistant; its plans page lists no video features and no bundled video app.
Claude makes no video on this plan (Claude Motion is marked No on Free, Pro and Max), so a video watermark does not apply.
GitHub Copilot is a coding assistant; its plans page lists no video features and no bundled video app.
Artifacts (all plans) turn a description into single-page websites, small interactive tools and apps that can call Claude; Claude Design makes prototypes and landing pages. Claude Code (included in paid plans) builds full apps in your codebase.
GitHub Spark, Copilot's app builder, is 'Not included' on this plan in the plans docs, and it stopped accepting new users and apps on August 4, 2026. GitHub points builders to Copilot in VS Code, Copilot CLI and the Copilot app, which are coding agents.
Artifacts start private; Pro and Max can share with 'Anyone with the link' (viewers sign in to Claude; not for artifacts that use Claude or connectors). Claude Code artifacts publish to a claude.ai URL that can be public with no sign-in; one self-contained page, no backend.
No hosting for new apps: GitHub Spark, which deployed apps, is not included on this plan and closed to new apps on August 4, 2026.
No Anthropic page documents serving an artifact on your own domain. Legacy published artifacts offer embed code with an Allowed domains list, which is embedding, not a custom domain.
No Copilot app hosting on this plan (GitHub Spark is not included and closed to new apps), so there is no domain to set.
Artifacts can store data between sessions: personal or shared storage, 20 MB per artifact, text only. Visitors of AI-powered artifacts sign in with their own Claude account. No other database or auth service is documented.
No Copilot app builder on this plan (GitHub Spark is not included and closed to new apps), so no built-in app storage.
Claude Design exports a .zip, standalone HTML, or hands off to Claude Code or tools such as Lovable, Replit, v0 and Vercel; Claude Code (included in paid plans) writes code in your own project, and published artifact pages come from a local file.
Copilot writes code into your own files and repos: agent mode edits your local project and Copilot cloud agent works in your GitHub repositories. Spark users could export an app to a repository until August 31, 2026.
No official Claude page describes generating songs or any audio output.
GitHub Copilot is a coding assistant and does not generate music.
No official Claude page describes generating instrumental music.
GitHub Copilot is a coding assistant and does not generate music.
No official Claude page describes stem downloads.
GitHub Copilot is a coding assistant with no music or audio output to split into stems.
Voice mode lets Claude talk back in a spoken conversation (pricing table: Voice mode Yes), but no official page describes turning text you supply into a voiceover file.
GitHub Copilot is a coding assistant and does not produce voiceovers from text.
No official Claude page describes copying a voice.
GitHub Copilot is a coding assistant and does not create synthetic voices.
No official Claude page describes translating and re-voicing audio or video.
GitHub Copilot is a coding assistant and does not translate or re-voice audio or video.
No official Claude page describes generating sound effects.
GitHub Copilot is a coding assistant and does not generate audio.
The upload help article lists document types (PDF, DOCX, CSV, TXT and more) and images only; no audio or video file formats are listed and transcription is not described.
GitHub Copilot is a coding assistant and does not transcribe recordings or audio and video files.
No official page describes Claude joining or recording meetings. Connectors (for example Microsoft 365) can read existing meeting notes or summaries.
GitHub Copilot is a coding assistant and does not join or record meetings.
Dictation is available to all plans, but only inside the Claude iOS and Android apps to speak prompts; no official page describes typing into other apps.
GitHub Copilot is a coding assistant and does not offer system-wide dictation into other apps.
Projects.
Copilot Spaces: collect repos, files and other sources as shared context for Copilot Chat on GitHub (and in IDE via GitHub MCP server); share privately, with users or publicly. Plus repository/personal custom instructions.
No GPTs/Gems-style assistants listed; Skills and Projects listed separately.
Custom agents: Markdown agent profiles (.github/agents/NAME.md) with prompt, tools and MCP servers; usable in cloud agent, CLI, VS Code, JetBrains, Eclipse, Xcode. Plans page ticks 'Custom instructions and agents' on all plans.
Memory across conversations.
Copilot Memory (public preview): stores repository facts and personal preferences across sessions, validated against code; unused entries deleted after 28 days; you can view/delete your preferences.
Directory connectors (e.g. Google Drive, Gmail, Google Calendar, GitHub, Slack, Microsoft 365) + custom remote MCP. Also Claude for Microsoft 365.
Model Context Protocol (MCP) server integration on all individual plans; GitHub MCP Server; cloud agent integrations with Slack, Teams, Jira, Linear, Azure Boards (paid).
No Claude add-on that works inside Gmail is documented. Claude in Chrome (a browser agent) can act on Gmail pages, and the Gmail connector works from Claude's own chat; neither is an integration inside Gmail.
GitHub's Copilot plans and features docs list IDEs, GitHub.com, GitHub Mobile, Windows Terminal, Copilot CLI and the Copilot app, plus Slack, Teams, Jira, Linear and Azure Boards integrations. No official GitHub page mentions an add-in inside Gmail.
Claude for Google Workspace add-on: Claude sidebar in Google Docs (Extensions > Claude). Beta; Chrome, Edge, or Safari (not Firefox); Google account must be allowed to install Marketplace apps.
GitHub's Copilot plans and features docs list IDEs, GitHub.com, GitHub Mobile, Windows Terminal, Copilot CLI and the Copilot app, plus Slack, Teams, Jira, Linear and Azure Boards integrations. No official GitHub page mentions an add-in inside Google Docs.
Claude for Google Workspace add-on: Claude sidebar in Google Sheets (Extensions > Claude). Beta; Chrome, Edge, or Safari (not Firefox); Google account must be allowed to install Marketplace apps.
GitHub's Copilot plans and features docs list IDEs, GitHub.com, GitHub Mobile, Windows Terminal, Copilot CLI and the Copilot app, plus Slack, Teams, Jira, Linear and Azure Boards integrations. No official GitHub page mentions an add-in inside Google Sheets.
Claude for Word add-in, generally available. Word on the web, Windows (Microsoft 365, Version 2205+), and Mac (16.61+). Pricing table: Claude for Microsoft 365 Yes.
GitHub's Copilot plans and features docs list IDEs, GitHub.com, GitHub Mobile, Windows Terminal, Copilot CLI and the Copilot app, plus Slack, Teams, Jira, Linear and Azure Boards integrations. No official GitHub page mentions an add-in inside Word.
Claude for Excel add-in, generally available. Excel on the web, Windows (Microsoft 365 subscription), and Mac (16.46+); not Excel on iPad. Pricing table: Claude for Microsoft 365 Yes.
GitHub's Copilot plans and features docs list IDEs, GitHub.com, GitHub Mobile, Windows Terminal, Copilot CLI and the Copilot app, plus Slack, Teams, Jira, Linear and Azure Boards integrations. No official GitHub page mentions an add-in inside Excel.
Claude for PowerPoint add-in, generally available. PowerPoint on the web, Windows (Microsoft 365 subscription), and Mac (16.46+). Pricing table: Claude for Microsoft 365 Yes.
GitHub's Copilot plans and features docs list IDEs, GitHub.com, GitHub Mobile, Windows Terminal, Copilot CLI and the Copilot app, plus Slack, Teams, Jira, Linear and Azure Boards integrations. No official GitHub page mentions an add-in inside PowerPoint.
Claude for Outlook add-in, Beta. Needs Exchange Online through Microsoft 365 and a one-time tenant-wide Microsoft Graph consent by a Global Administrator; not Outlook on iOS or Android. Pricing table: Yes.
GitHub's Copilot plans and features docs list IDEs, GitHub.com, GitHub Mobile, Windows Terminal, Copilot CLI and the Copilot app, plus Slack, Teams, Jira, Linear and Azure Boards integrations. No official GitHub page mentions an add-in inside Outlook.
Claude in Chrome; hand off and schedule tasks.
Copilot agents work in your editor, terminal or a cloud dev environment; no browser- or computer-operating agent documented on the pages checked.
User chooses 'Model Improvement' at signup/in-app; changeable anytime in Privacy Settings. Default not stated.
Since Apr 24 2026 GitHub may train on Free/Pro/Pro+/Max interaction data by default; you can opt out anytime ('Allow GitHub to use my data for AI model training' -> Disabled).
Export from web or desktop (not iOS/Android).
Plans FAQ only points to the GitHub Privacy Statement for access/deletion; no Copilot chat-history export documented.
Ad-free commitment.
Advertising not mentioned on the pages checked.
Conflict: Anthropic's help center says you own outputs and may use them in your products; the Consumer Terms we could read say non-commercial use only. Check the terms for your account.
Not stated as a permission: GitHub's Terms (Section J) say GitHub does not claim ownership of Output and set no commercial-use ban; no page grants commercial use in so many words.
No cloud-storage bundle.
No cloud storage bundled with Copilot is documented.
Web, iOS, Android, desktop.
IDE extensions (VS Code, Visual Studio, JetBrains, Xcode, Eclipse, Neovim, Zed, SSMS, Raycast), Copilot CLI, GitHub Copilot app (desktop), github.com and GitHub Mobile. Docs: all plans include CLI and Copilot app.
Copilot Pro is $10 a month. GitHub sells it monthly to new subscribers; annual Pro plans are legacy, and existing annual subscribers move to Free when theirs expire. Claude Pro is $20 a month, or $200 a year on the web paid up front, about $17 a month. Neither vendor says on the pages we checked whether US sales tax is included, so all prices here are before tax.
The next steps up differ. Copilot Pro+ is $39 a month and Copilot Max $100. Claude Max 5x is $100 and Claude Max 20x $200, both monthly only.
GitHub documents Copilot extensions for VS Code, Visual Studio, JetBrains IDEs, Xcode, Eclipse, Neovim, Zed, SQL Server Management Studio and Raycast, with agent mode in VS Code, Visual Studio, JetBrains IDEs, Eclipse and Xcode. Every plan includes Copilot CLI and the Copilot desktop app. Features vary by editor.
Claude Code, from Pro up, documents an extension for VS Code and VS Code forks such as Cursor and Kiro, and a plugin for JetBrains IDEs including IntelliJ IDEA, PyCharm, Android Studio, WebStorm, PhpStorm and GoLand. Visual Studio, Xcode, Eclipse, Neovim and Zed are not listed in the Claude Code facts we checked, so support there is not verified.
On Copilot Pro, inline suggestions and next edit suggestions are unlimited and do not use AI credits; Copilot Free caps completions at 2,000 a month. Anthropic describes Claude Code in VS Code and JetBrains as an agent with a chat panel and inline diffs. No Anthropic page we checked documents completions as you type, and the Desktop app docs say it has no autocomplete, so for Claude this is not verified rather than documented as not offered.
Both document an agent that edits your code in the editor: Copilot's agent mode uses AI credits, and Claude Code shares Pro's limits with chat. Whether the changes work is for testing: our code repair kit and code refactor kit give both the same task.
Copilot's cloud agent researches a repository, plans and codes in a cloud environment, then opens a pull request. It works only on GitHub-hosted repos, handles one repo, branch and pull request per task, stops at a 59-minute session limit and uses AI credits. Claude Code on the web runs cloud sessions on Anthropic-managed virtual machines that keep running after you close your laptop; they clone GitHub repos, create pull requests and can auto-fix pull requests. There is no separate machine charge, and sessions count toward Pro's usage limits.
For review, Copilot Pro reviews GitHub pull requests on request or automatically for pull requests you create, and reviews file diffs in the editor. GitHub puts a review at about $0.05 to $1 of credits in Lite mode, or $0.25 to $5 in Balanced mode. On Claude Pro, review runs through Claude Code GitHub Actions authenticated with your subscription token, and the review workflow posts inline comments automatically. Anthropic's managed Code Review, a research preview, is for Team and Enterprise only. Neither vendor documents how useful its reviews are; hand each agent the same task from the code spec kit.
Claude Pro lists only Anthropic models: Opus, Sonnet and Haiku, with Fable through usage credits. Copilot Pro lists models from Anthropic, OpenAI, Google, xAI, Moonshot and Microsoft, paid from AI credits, plus an auto mode that takes 10% off model cost. Copilot Free uses auto model selection only, and GitHub does not list which models that uses.
Claude models appear in Copilot by plan. Pro lists Claude Sonnet 5.5 and Haiku 5.5; Claude Opus 5.5 and Claude Fable 5.1 are listed for Pro+ and Max. Pro+ and Max can also delegate tasks to third-party agents, Claude and Codex. Anthropic documents up to 1 million tokens of context on all individual plans; GitHub offers 1 million as an option for supported models in VS Code and Copilot CLI only, at a higher credit cost.
Copilot Pro gives 1,500 GitHub AI Credits a month, 1,000 base plus 500 flex, at $0.01 each; GitHub says the flex part may change. Chat, agent mode, code review, the cloud agent and the CLI use credits; completions do not. Past the allowance you pay from a dollar budget you set. Pro+ includes 7,000 credits a month and Max 20,000.
Claude Pro has a session limit that resets every five hours plus a weekly limit, and chat and Claude Code draw from the same pool. Anthropic puts Pro at least five times Free usage per session; Max 5x and Max 20x give five and twenty times Pro. Apart from Fable's usage credits, the facts we checked do not describe buying extra usage on Pro, so the documented path for heavy use is Max.
Claude Pro is also a general assistant: Research reports, web search, file uploads of up to 500MB each, voice mode, memory, projects, connectors such as GitHub and Google Drive, and Claude in Chrome. It does not generate photos or illustrations. GitHub scopes Copilot to code, repositories and development concepts; image generation and spoken conversation are not documented for it.
GitHub says that since April 24, 2026 it may train on interaction data from Free, Pro, Pro+ and Max by default; set "Allow GitHub to use my data for AI model training" to Disabled to opt out. Anthropic asks you to choose the Model Improvement setting at sign-up and does not state the default, so check it. Anthropic states an ad-free commitment; advertising is not mentioned on the Copilot pages we checked.
Copilot Pro. GitHub documents unlimited inline and next edit suggestions on Pro that do not use AI credits. For Claude Code, completions as you type are not documented on any plan, so that is not verified.
Copilot Pro. GitHub documents extensions for all four, with agent mode in Xcode, Eclipse and Visual Studio. The Claude Code facts we checked list VS Code, VS Code forks and JetBrains IDEs only, so other editors are not verified.
Claude Pro includes Opus inside its five-hour and weekly limits, through Claude Code. In Copilot, Claude Opus 5.5 is listed only on Pro+ ($39 a month) and Max ($100); Copilot Pro lists Claude Sonnet 5.5 and Haiku 5.5.
Claude Pro documents cloud sessions that keep running after you close your laptop, with no separate machine charge. Copilot's cloud agent stops at a 59-minute session limit and handles one repo, branch and pull request per task. Test both with the code spec kit.
Claude Pro. It includes Research, web search, file uploads, voice and Claude in Chrome alongside Claude Code, in one usage pool. GitHub scopes Copilot to software work, so it is not built as an everyday assistant.
Both document it. Copilot Pro reviews pull requests you create automatically, at about $0.05 to $5 of credits per review. Claude Pro uses a GitHub Actions review workflow with your subscription token, which posts inline comments. Whether the reviews catch real bugs is untested.
We have not run these tests for you unless a result above says so. Each kit gives you the same files, the same prompt and the same checks for both tools.
Yes, by plan. Copilot Pro lists Claude Sonnet 5.5 and Haiku 5.5. Pro+ and Max add Claude Opus 5.5 and Claude Fable 5.1, and can delegate tasks to a Claude agent. Copilot Free uses auto model selection only, and GitHub does not list which models that uses. Model use comes out of Copilot's AI credits.
Not verified. Anthropic documents Claude Code in VS Code and JetBrains as an agent with a chat panel and inline diffs, and no page we checked documents completions as you type. Copilot Pro documents unlimited inline suggestions; Copilot Free allows 2,000 a month.
Copilot Pro meters use in AI credits: 1,500 a month, $0.01 each, with completions free of credits and extra use billed against a budget you set. Claude Pro has a five-hour session limit and a weekly limit, shared between chat and Claude Code, with Max for five or twenty times Pro's usage.
On Copilot, yes: Free includes 2,000 completions a month and limited chat and agent use, though GitHub's plans FAQ still says 50 chat requests a month, so the chat allowance is not verified. Claude Free includes Claude chat but not Claude Code, so no editor integration or cloud sessions.
GitHub says that since April 24, 2026 it may train on interaction data from Copilot Free, Pro, Pro+ and Max by default, and you can opt out at any time in settings. Anthropic asks you to choose the Model Improvement setting at sign-up and does not state which option is the default.
Neither vendor's documentation answers that, and we have not run our tests on these plans yet. Give both the same task from our code repair, code spec or code refactor kits, in the same repository, and compare the results against the kit's checks.
Every fact links to its source and shows when it was verified. Read how AI Compare works and the change log.