If cost is the reason, check Copilot's effort level first: GitHub estimates a review at $0.05 to $1 worth of AI credits on Lite and $0.25 to $5 on Balanced, its default since 28 September 2026, not counting Actions minutes. Documented alternatives include Claude Code, CodeRabbit, Codex from OpenAI, Cursor Bugbot, Graphite, Greptile and Qodo, each billed in its own unit. With no budget, one developer can start on Greptile Starter, Graphite Hobby or CodeRabbit's Free plan in the IDE and CLI; in a subscription you may already have, /code-review works on any plan that includes Claude Code, and ChatGPT Plus includes Codex code review. Choose by how the bill grows (per developer, per review or run, or a credit pool), where your code lives, and who pays.
Every fact below comes from each vendor's own pages as I read them on 23 October 2026, and none of the vendors saw this text before publication.
Copilot code review alternatives side by side
Each cell is from the vendor's own pages, read on 23 October 2026; "not found" means not found on the pages listed at the end that day.
| Tool | Code hosts | When it reviews by default | What you pay for | Free or trial entry |
|---|---|---|---|---|
| GitHub Copilot code review | GitHub.com, GitHub CLI, GitHub Mobile and IDEs; Azure DevOps (public preview); "Copilot is not currently available for GitHub Enterprise Server." | when someone assigns it; automatically by an author's own setting (Pro, Pro+, Max, Business, Enterprise) or a ruleset; once per pull request unless set to review each push or asked again | AI credits from your plan, estimated per review at $0.05 to $1 on Lite and $0.25 to $5 on Balanced, plus Actions minutes on GitHub-hosted runners | Copilot Free: no pull request review, only "Review selection" in VS Code; Copilot Student for verified students; Pro at $10 USD per month, "free for some users" |
| Claude Code (Anthropic) | managed Code Review: GitHub, including GitHub Enterprise Server and GitHub Enterprise Cloud with data residency; /code-review --comment: a GitHub pull request or a GitLab merge request; GitHub Action: GitHub; GitLab CI/CD: GitLab, "currently in beta" |
Code Review: per repository, Once after PR creation, After every push or Manual, and a repository added from the admin page starts at After every push; /code-review: when you run it; GitHub Action or GitLab CI/CD: your pipeline's events |
Code Review, "out of research preview" since 15 October 2026 for Team and Enterprise: token usage from usage credits, billed to the organization, "no flat per-review price"; /code-review: your plan's usage or API billing; GitHub Action: an API key, cloud provider credentials, or a Pro, Max, Team or Enterprise subscription token, plus Actions minutes |
ultrareview: 3 free runs on Pro and Max, once per account, then "typically $5 to $25 in usage credits depending on the size of the change"; managed Code Review: "Eligible Claude Enterprise organizations can claim at least $10,000 in Code Review credits by January 15, 2027.", and the credits expire 30 days after the claim |
| CodeRabbit | GitHub (including GitHub Enterprise Server), GitLab (including self-managed), Azure DevOps, Bitbucket Cloud and Bitbucket Data Center; self-hosted on Enterprise from 500 user seats | every eligible pull request to the default branch, skipping drafts, then later pushes; @coderabbitai review on request |
each developer who creates pull requests: Essentials $24 per developer per month billed annually, $30 monthly; 5 PR reviews per hour per developer, then $0.25 per reviewed file only with usage-based reviews turned on | Free: pull request summaries, with reviews in the VS Code extension and CLI; public repositories "receive free reviews forever"; "New organizations start with a 14-day Advanced trial" |
| Codex (OpenAI) | GitHub, where OpenAI's Code review page says "GitHub code review is generally available."; GitLab merge requests, in preview by that page and in beta by the GitLab page | on request with @codex review; automatically once a person turns on Automatic review in their own settings; in GitHub it flags only P0 and P1 issues |
your ChatGPT plan's usage, Plus at $20 a month and higher; reviews through GitHub count as the plan's Code Review usage; not with an API key | GitHub code review was not found among the Free and Go plan features on OpenAI's pricing pages as of 23 October 2026; the GitLab page says "GitLab support is in beta and is available on all ChatGPT plans." |
| Cursor Bugbot | Origin, GitHub (including GitHub Enterprise Server), GitLab (including GitLab Self-Hosted), Bitbucket (including Bitbucket Data Center), Azure DevOps Services | "Bugbot runs automatically when a PR is created or updated."; a personal setting runs it only when mentioned or only once per pull request | each run from your Cursor usage, on average $1.00 to $1.50 per run by Cursor's help page; Premium at $30 a month is the first plan with Bugbot, Teams at $30, $100 or $200 per seat a month | a free trial was not found on the Bugbot and pricing pages as of 23 October 2026; "Enterprise teams can arrange a Bugbot trial through their account team." |
| Graphite | GitHub; GitHub Enterprise Server on Enterprise | every new pull request in the repositories you select | per user per month billed annually: Starter $20, Team $40; AI reviews limited on Hobby and Starter, unlimited on Team | Hobby is free for "Personal account repos" with limited AI reviews, no number on the pricing page; "Free for your first 30 days. No credit card required." |
| Greptile | GitHub, GitLab, Bitbucket, Cursor Origin; on Enterprise also GitHub Enterprise Server, GitHub Enterprise Cloud with data residency, GitLab Self-Managed, Bitbucket Data Center and Gitea; for on-premises deployments, Perforce through P4 Code Review | when a pull request opens (autoReview defaults to ["open"]); @greptileai on request |
Pro: $1 per credit with no seat fees, credits shared across the organization; a Base review is 1 credit, Plus 3, Apex 10 | Starter: one active developer, unlimited repositories, 50 credits per month; a 14-day Pro trial; free for qualified non-commercial projects; a startup discount of 500 free credits a month for 6 months for pre-Series A startups with under $2M revenue |
| Qodo | GitHub (cloud and Enterprise Server), GitLab (cloud and self-managed), Bitbucket (Cloud and Data Center), Azure DevOps; Gerrit on Enterprise | when a pull request is opened, reopened or marked ready for review; /agentic_review on request |
a credit pool for the team: Pro Team $30 a month for 2,500 credits, ~18 reviews by Qodo's estimate, "$.012/credit, pooled across the team"; "Designed for up to 30 users" | "Free 14 Day Trial"; "We don't offer a permanent free tier."; Qodo for Open Source by application |
| Sigma (Mnemoverse, the author's company) | GitHub: "Careful pull request reviews in GitHub." | invited accounts start with On request: comment /sigma review on a pull request; documented modes are Automatic, On request and Off |
Closed beta, by invitation | Closed beta; "Invited GitHub accounts can sign in." |
If the bill is the reason
GitHub's docs say "Code reviews have two cost components": AI credits for the review itself and GitHub Actions minutes for context gathering and tool use, except on self-hosted runners. GitHub's own advice on effort is "Use Lite for routine changes where fast feedback is more important than exhaustive analysis." By default a review the repository requests automatically is billed to the pull request author, and a manual one to whoever asked. Since 8 October an organization owner can bill licensed members' reviews to the organization instead, but the member stays the default, and "If the member's quota is exhausted, the code review fails."
The alternatives count in other units. CodeRabbit charges only for "developers who create pull requests", Graphite per user, and Qodo for a credit pool the team shares, while Greptile writes "Pro has no seat fees: you pay $1 per credit, and credits are shared across your whole organization." Cursor bills each Bugbot run from the same usage as the editor, "The average Bugbot run costs between $1.00 and $1.50, depending on PR size and complexity.", and a pull request gets a run on every update unless set to once. Cursor's $8 Plus plan excludes Bugbot, and its docs bill reviews past the included usage "from your prepaid credit, or from on-demand spend on a legacy Cursor plan", while its help page still describes on-demand usage. Anthropic bills managed Code Review on tokens to the organization; OpenAI counts Codex reviews through GitHub as Code Review usage on the ChatGPT plan.
Two lines of arithmetic on the vendors' own numbers, not measurements: Greptile Starter's 50 credits a month cover up to 50 Base reviews at 1 credit each, 16 Plus reviews at 3 credits or 5 Apex reviews at 10; Qodo's $30 for what Qodo estimates as ~18 reviews comes to about $1.67 a review. Which bill is lower depends on your pull requests and pushes, so price last month's in each vendor's unit before you switch.
One developer, private repositories
This is the case behind the question: one developer, private GitHub repositories, little budget. Greptile's pricing page says "Starter is free for one developer and includes 50 credits per month." and its Starter card lists "Unlimited repositories". Graphite's free Hobby plan covers "Personal account repos" with "Limited AI Reviews". The word private was not found on Greptile's pricing page as of 23 October 2026, and Graphite's pricing page uses it only for an Enterprise feature. CodeRabbit's Free plan names them, "Unlimited public and private repositories, no credit card required.", with reviews in the VS Code extension and CLI.
If you already pay for a subscription, look there. In Claude Code, /code-review "works on any plan" within your plan's usage, and --comment posts its findings on a GitHub pull request. The Claude Code GitHub Action can use a Pro or Max subscription: "If you authenticate with an OAuth token, runs use your Claude subscription instead of API billing.", plus Actions minutes. Ultrareview, "a research preview feature", gives Pro and Max three free runs, "a one-time allotment per account".
ChatGPT Plus, $20 a month on OpenAI's Codex pricing page, includes "Cloud-based integrations like automatic code review and Slack integration"; you ask with @codex review. On an API key, OpenAI lists "No cloud-based features (GitHub code review, Slack, etc.)". On Cursor Premium, Bugbot draws on the plan's usage, and on an individual plan "Bugbot runs only on PRs you author." On GitHub, Copilot Free "does not include Copilot code review", while Copilot Student, "Available to verified students. Get access to Copilot's features for free.", has code review marked as included in GitHub's plans table.
If you came from Gemini Code Assist on GitHub, Google's deprecation page says "the consumer version is shut down, and all code review activities performed by the app end." and "The enterprise version of Gemini Code Assist on GitHub is not affected by these changes." That version is in Preview, and Google writes "there are no charges for using Gemini Code Assist on GitHub during Preview" and "While a valid billing account is required, it only accrues charges when usage goes over the free tier limits."
What happens before a comment reaches you
Each vendor describes its own filter, and none of these descriptions is a measurement. CodeRabbit's architecture page lists "Specialized AI agents working in parallel: Review, Verification, Chat, Pre-Merge Checks" and "Sandboxed cloud execution with your full repository cloned for isolated analysis". Anthropic's Code Review docs say "A verification step checks findings against the code before they're reported." and that a review works "without building or running your code"; for ultrareview, "every reported finding is independently reproduced and verified". Greptile's homepage says "TREX runs your PR branch in a sandbox to catch the bugs that only show up at runtime."; its docs changelog says "TREX is generally available" in an entry dated 16 October 2026 and "A TREX run costs 3 credits and works with every review tier.", and its TREX docs add that an organization turns TREX on in its settings and that a run is billed apart from the code review. Qodo describes "filtering out low-impact or irrelevant comments before they reach the pull request". Codex limits its GitHub comments to P0 and P1 issues, and Cursor says "Bugbot uses a combination of frontier and in-house models to review code." GitHub says "Copilot code review runs in an ephemeral development environment, similar to Copilot cloud agent." and also "Copilot is not guaranteed to spot all problems or issues in a pull request. Sometimes it will make mistakes." For published accuracy numbers and their methods, see Should we trust AI code review benchmarks?.
Sigma, in closed beta on GitHub, says on its homepage that it "checks potential issues and brings the evidence back to your code"; its docs say a fresh verifier "confirms, rejects or reclassifies each one", that "only confirmed candidates are verified findings", and that suspicions no verifier confirmed stay labeled separately. Sigma runs on Claude Code and OpenAI models and reviews its team's own repositories every day. Its homepage takes requests through a "Request beta access" form and says "A request does not grant access automatically."
How to choose
| If this is true for you | Look first at | What the pages say |
|---|---|---|
| One developer, private repositories, no budget | Greptile Starter, Graphite Hobby, CodeRabbit in the IDE and CLI | Starter: one active developer, 50 credits a month, a Base review is 1 credit. Hobby: personal account repositories, limited AI reviews. CodeRabbit Free: "Unlimited public and private repositories", reviews in the VS Code extension and CLI |
| You pay for Copilot and the bill is the problem | Copilot at Lite for routine pull requests | GitHub's own advice for Lite; since 8 October an owner can bill licensed members' reviews to the organization |
| You pay for a Claude plan |
/code-review; the GitHub Action with a subscription token; managed Code Review on Team and Enterprise |
/code-review on any plan within its usage; the Action on Pro, Max, Team or Enterprise, plus Actions minutes; Code Review billed to the organization |
| You pay for ChatGPT Plus or higher | Codex code review | reviews through GitHub count as the plan's Code Review usage; only P0 and P1 issues in GitHub; not on an API key |
| You pay for Cursor Premium or Teams | Bugbot | runs from the same Cursor usage, $1.00 to $1.50 per run on average by Cursor's help page, one run per update unless set to once |
| Code on GitLab, Bitbucket or Azure DevOps | CodeRabbit, Cursor Bugbot, Qodo; Greptile for GitLab and Bitbucket; on GitLab also Claude Code (/code-review --comment, GitLab CI/CD in beta) and Codex |
the first three list all three hosts; Azure DevOps was not found among the code providers Greptile lists on its homepage and code-providers page as of 23 October 2026; GitLab, Bitbucket and Azure DevOps were not found on Graphite's docs index or pricing page that day |
| GitHub Enterprise Server | CodeRabbit, Cursor Bugbot; Claude's managed Code Review; Greptile, Graphite and Qodo on Enterprise | "Copilot is not currently available for GitHub Enterprise Server."; Qodo's deployment table marks "GitHub Enterprise", listed apart from "GitHub Cloud Enterprise", for single-tenant, on-prem and air-gapped deployments and not for multi-tenant, and its pricing page includes single-tenant and on-prem on Enterprise |
| Code must stay on your own servers | CodeRabbit Enterprise from 500 user seats, Greptile Enterprise (air-gapped too), Qodo Enterprise (on-prem or air-gapped) | "There is no self-hosted deployment of Graphite's service available for a user to run in their own cloud."; managed Code Review runs on Anthropic infrastructure, while for Claude in GitHub Actions or GitLab CI/CD Anthropic's table lists "Your own CI runners" with a Claude API key or cloud provider credentials; Copilot's review workflow can use self-hosted runners, but a self-hosted Copilot or Bugbot review service was not found on GitHub's code review page or Cursor's Bugbot docs as of 23 October 2026 |
| The bill should follow headcount, not review volume | CodeRabbit, Graphite | per developer or per user; Copilot, Bugbot and managed Code Review count reviews, runs or tokens; Codex counts plan usage; Greptile Pro sells credits with no seat fees; Qodo pools credits |
| Each finding should pass a separate check before it is posted | CodeRabbit; Claude's managed Code Review and ultrareview; Qodo | a Verification agent; a verification step; findings "independently reproduced and verified"; a filter before comments reach the pull request |
Run two of them on the same pull requests
Per-review prices here are estimates and averages, and each filter is the vendor's own description, so test on your own code. Pick a private repository and three recent pull requests where a bug surfaced after merge. Branch from the commit before each fix, open it again as a new pull request, and ask two candidates to review it, for example Copilot at Lite and one tool from the table that fits your budget. For each tool, note whether it named the known bug, how many comments you would have dismissed, and what the review cost in its own unit, from its usage page. Three pull requests will not rank the tools, but they show whether the bill and the noise are acceptable.
Other tools also review pull requests. Devin's Free plan includes access to Devin Review, its billing page says "Public PRs are free.", and its docs list GitHub (including GitHub Enterprise Server and Enterprise Cloud), GitLab (including Self-Managed GitLab), connected Azure DevOps repositories and Bitbucket Data Center; when credits run out, it "switches to its smart diff viewer".
Head-to-head pages, each describing vendor pages as of the read date it gives: CodeRabbit vs Copilot?, Qodo vs CodeRabbit?, CodeRabbit vs Greptile?.
Sources, all read on 23 October 2026; the text describes these pages as of that day. GitHub: Copilot code review, configuring code review, using code review, plans (docs), plans and FAQ, effort default, organization billing. Anthropic: Code Review, setup, costs, /code-review, ultrareview, GitHub Actions, GitLab CI/CD, general availability. CodeRabbit: pricing, plans, rate limits, usage-based reviews, platforms, automatic reviews, self-hosted, architecture. OpenAI: Codex code review, GitHub, GitLab, pricing (docs), Codex pricing. Cursor: Bugbot, Bugbot docs, Bugbot help, Bugbot billing, models and pricing, pricing. Graphite: pricing, homepage, AI reviews setup, GitHub Enterprise Server, docs index. Greptile: pricing, homepage, code providers, config reference, self-hosting, changelog, TREX. Qodo: pricing, homepage, triggering a review, how review works, deployment models. Google: consumer deprecation, enterprise version, setup. Cognition: Devin Review, billing. Sigma: homepage, reading a review, modes and requests.
Top comments (0)