by GitHub, Inc.AI coding assistants
An AI coding assistant for many editors with unlimited completions, chat, agents and pooled AI credits: Business at $19 per user a month.
Unlimited completions plus chat and agents on models from several providers, with no training on your code and IP indemnity from $19 a user.
Copilot Business, at $19 per user a month, gives each developer unlimited code completions and next edit suggestions, and 1,900 AI credits ($19 of chat and agent usage) a month pooled across the organization, with models from Anthropic, OpenAI, Google, xAI and others. GitHub states that Business data is not used to train its models, offers IP indemnity for unmodified suggestions when the public-code filter is on, and lists SAML SSO, an audit log, content exclusion and budgets on Business.
Against 3 other products we reviewed in AI coding assistants: what you pay for with each, the five access facts, the plan limits we recorded and each review's bottom line. The note under the table says how each one was checked.
| Metric | GitHub Copilot | Cursor | Claude Code | JetBrains AI |
|---|---|---|---|---|
| What you're paying for | Compliance includedUnlimited completions plus chat and agents on models from several providers, with no training on your code and IP indemnity from $19 a user. | Saves engineering timeAn editor built around coding agents, with zero data retention with model providers and privacy mode enforced for the team on Teams. | Saves engineering timeA coding agent for the terminal, IDEs and the web, sold with Claude chat on one seat, with no training and IP indemnity on Team. | Worth the priceAI completion, chat and agents inside JetBrains IDEs, with usage at API rates and no surcharge, and free governance for organizations. |
| Overall score | 8.3Best | 6.4 | 5.6 | 5.3 |
| Free option | Free plan | Free plan | Paid only | Free plan |
| Card to start | –Not measured | –Not measured | –Not measured | Card required |
| Getting access | Self-serve | Self-serve | Self-serve | Self-serve |
| Starter priceCheapest paid plan | $10/mo | $8/mo | $20/mo | $10/mo |
| Pricing model | Seat + usage | Seat + usage | Per seat | Seat + usage |
| Pricing page | Public prices | Public prices | Public prices | Public prices |
| Billing unit | Per user per month; chat and agent usage in AI credits ($0.01 each) | Per seat per month; third-party models from prepaid credit at API list price | Per seat per month; extra usage as usage credits at API rates | AI credits ($1 of use each); individual plans per month; governance per seat |
| Free option | Copilot Free: 2,000 completions a month, limited chat and agent usage | Free plan with limited Agent and Chat on Cursor models; Tab on Free not stated | None with Claude Code; the Free plan excludes it | AI Free with a JetBrains IDE licence; 30-day AI Pro trial, card needed |
| Reference scenario (5 developers) | $95/mo on Business ($19 × 5, billed monthly), computed | $150/mo on Teams Premium seats ($30 × 5, monthly), computed | $100/mo on Team Standard seats ($20 × 5, annual), computed | Not priced: no organization plan includes usage (credits bought separately) |
| Included usage (organization plan) | 1,900 AI credits ($19) per user a month, pooled; completions unlimited | Premium weekly pool per seat (4x Plus); no absolute size stated | 1.25x Pro per session per seat, shared with chat; no absolute size stated | None; pooled prepaid credits, 20 per seat with All Products Pack |
| Usage beyond the plan | Per-token rates in AI credits; on by default for organizations, budgets available | Prepaid credit at per-token rates, plus $0.25 per million tokens on Teams | Usage credits at API rates, off until turned on; spend limits | Prepaid credits at API rates, no surcharge; per-user limits |
| Models | Anthropic, OpenAI, Google, xAI, Microsoft and Moonshot models on Business | Cursor (Grok, Composer), Anthropic, OpenAI and Google models | Anthropic only (Fable, Opus, Sonnet, Haiku) | Anthropic, OpenAI and Google models; JetBrains Mellum for completion |
| Training on your code | Not on Business and Enterprise; Free to Max train unless you opt out | Not without explicit agreement (terms); Privacy Mode enforced on Teams | Not on Team and Enterprise by default; Free, Pro and Max opt-out | Not without your agreement (terms), on every tier |
| Data retention | IDE prompts not retained; other surfaces 28 days (Business) | Zero data retention with providers in Privacy Mode, with exceptions | 30 days on Team; zero data retention on qualified Enterprise accounts | Not persisted unless detailed data collection is on; provider retention not stated |
| IP indemnity | Business and Enterprise, unmodified suggestions with the filter on | In the MSA; plans covered not stated | Commercial plans (Team, Enterprise, API), no liability cap | Not offered |
| SSO / SCIM / audit log | SAML SSO and audit log listed for Business; SCIM through managed users in an enterprise account | SAML SSO on Teams; SCIM and audit log on Enterprise | SAML SSO on Team; SCIM and audit logs on Enterprise | SSO listed on Basic; SAML in IDE Services; SCIM coming soon; console audit page |
| SOC 2 / ISO 27001 | SOC 2 Type II and ISO 27001 cover Copilot Business and Enterprise | SOC 2 Type II, ISO 27001 and ISO 42001 | SOC 2 Type II, ISO 27001 and ISO 42001 | SOC 2 Type II covers JetBrains AI; ISO 27001 not listed |
| Editors and surfaces | VS Code, Visual Studio, JetBrains, Xcode, Eclipse, Neovim, Zed, CLI, github.com | Cursor app, CLI, web, iOS; JetBrains and Xcode integrations on paid plans | Terminal, VS Code, JetBrains, desktop, web, iOS and Android, Slack, CI | JetBrains IDEs, Android Studio, JetBrains Air app and web |
| Bottom line | From its public pages, Copilot Business puts unlimited completions, $19 of pooled usage per user, a wide model choice and the main enterprise controls on a $19 seat. Set budgets early: extra usage is on by default for organizations. | From its public pages, Cursor gives a team an agent-first editor with privacy mode enforced and SSO for $30 a seat. Included usage is stated only as multiples, several pages disagree, and SCIM and the audit log need Enterprise. | From its public pages, Claude Code gives a team an agent for the terminal, IDE and web, plus Claude chat, on a $20 annual seat with no-training terms and uncapped IP indemnity. Usage is stated only as multiples, and models are Anthropic-only. | From its public pages, JetBrains AI suits teams in JetBrains IDEs that want unlimited completion and usage at API rates with per-user limits. There is no IP indemnity, no seat with usage included, and several governance features are still coming. |
Evidence: GitHub Copilot: facts checked 2026-10-16 from its public sign-up and pricing pages, no completed hands-on run; Cursor: facts checked 2026-10-16 from its public sign-up and pricing pages, no completed hands-on run; Claude Code: facts checked 2026-10-16 from its public sign-up and pricing pages, no completed hands-on run; JetBrains AI: facts checked 2026-10-16 from its public sign-up and pricing pages, no completed hands-on run. Prices in USD as listed, excluding tax. "–" means not recorded.
Review from public pages
Written from GitHub Copilot's public pages (pricing, security or privacy, and docs) as read on Oct 16, 2026. We did not sign up, and no hands-on run is published yet.
We read GitHub's Copilot plans page and FAQ, the GitHub Docs articles on plans, licences, usage-based billing, models and pricing, budgets, seat assignment, content exclusion, model hosting and data residency, the Terms of Service, the Generative AI Services Terms, the Data Protection Agreement, the subprocessor list and GitHub's changelog posts on its SOC and ISO reports on Oct 16, 2026; we did not sign up, start a trial or buy a seat. The pages showed US dollar prices.
Billing changed this year. Since June 1, 2026, chat, agent and code review usage is billed in GitHub AI Credits (1 credit = $0.01) at per-token rates by model instead of premium requests; legacy annual subscribers keep the older model until their plan ends, and annual billing is no longer sold. Business includes 1,900 credits per user a month, pooled at the billing entity, and unused credits do not roll over.
Some pages disagree. The plans FAQ still limits Free to 50 chat requests, while the plans page and docs only say "limited"; the plans page lists the full model catalog on Free, while the docs say auto model selection only; github.com/pricing offers a 30-day trial that includes Copilot, while the Enterprise Cloud trial docs exclude Copilot Business and Enterprise; and the plans FAQ ties codebase indexing to Enterprise, while the indexing docs list no plan limit.
Code completions and next edit suggestions do not draw on AI credits on paid plans; Free has 2,000 completions a month.
vs Claude Code: Cursor does not say whether Tab is unlimited on its current plans; Claude Code has no inline completion.
One AI credit is $0.01, Business includes 1,900 credits per user a month pooled across the organization, and per-token rates are listed for each model, with 10% off model costs under auto model selection.
vs Claude Code: Cursor and Claude Code state their included usage only as multiples of another plan.
Business gets the full model catalog from Anthropic, OpenAI, Google, xAI, Microsoft and Moonshot, in VS Code, Visual Studio, JetBrains IDEs, Xcode, Eclipse, Neovim, Zed, the CLI and github.com.
vs JetBrains AI: Claude Code runs Anthropic models only; JetBrains AI runs in JetBrains IDEs.
The plans page lists IP indemnity and SAML SSO on Business, the plan docs add audit logs, content exclusion and organization-wide policy management, and SOC 2 Type II and ISO 27001 cover Copilot Business.
vs Claude Code: Cursor and Claude Code keep SCIM and the audit log on Enterprise.
Usage beyond the pooled credits is enabled by default for organizations, and organization budgets stop usage only when "Stop usage when budget limit is reached" is turned on, which is off by default. User-level budgets always stop usage.
Workaround: Set an organization budget and turn on the stop setting before assigning seats.
Only monthly prices are published and annual plans are no longer sold. Seats removed mid-cycle are billed to the end of the cycle without a refund, and removing seats does not shrink that month’s credit pool.
On Business, IDE chat and completion prompts are not retained; prompts from other surfaces are kept for 28 days. Claude Fable models are not under zero data retention by default.
Content exclusion is not supported in VS Code edit and agent modes, and IP indemnity applies only to unmodified suggestions with the public-code filter enabled.
Copilot bills each seat per month. For our scenario, 5 developers on Business cost 5 × $19 = $95 a month ($1,140 a year, computed), billed monthly because no annual price is sold. That includes unlimited completions and 1,900 AI credits ($19) per user, pooled into 9,500 credits ($95) a month. Beyond that, usage is billed at per-token rates, for example $4 in and $20 out per million tokens for Claude Opus 5.5, and data-residency or FedRAMP requests cost 10% more. Enterprise ($39) needs GitHub Enterprise Cloud, from $21 per user for the first 12 months, with the later price not stated. Code review also uses GitHub Actions minutes.
There is no code index to move. Usage metrics download as NDJSON through the REST API, the audit log keeps 180 days and can stream to a SIEM, and CLI and app sessions can be shared as a link, gist or file. There is no bulk admin export of chats, and cloud agent sessions cannot be deleted. Organization customers can migrate data for up to 90 days after the term.
From its public pages, Copilot Business puts unlimited completions, $19 of pooled usage per user, a wide model choice and the main enterprise controls on a $19 seat. Set budgets early: extra usage is on by default for organizations.
| Limit | As recorded |
|---|---|
| Billing unit | Per user per month; chat and agent usage in AI credits ($0.01 each) |
| Free option | Copilot Free: 2,000 completions a month, limited chat and agent usage |
| Reference scenario (5 developers) | $95/mo on Business ($19 × 5, billed monthly), computed |
| Included usage (organization plan) | 1,900 AI credits ($19) per user a month, pooled; completions unlimited |
| Usage beyond the plan | Per-token rates in AI credits; on by default for organizations, budgets available |
| Models | Anthropic, OpenAI, Google, xAI, Microsoft and Moonshot models on Business |
| Training on your code | Not on Business and Enterprise; Free to Max train unless you opt out |
| Data retention | IDE prompts not retained; other surfaces 28 days (Business) |
| IP indemnity | Business and Enterprise, unmodified suggestions with the filter on |
| SSO / SCIM / audit log | SAML SSO and audit log listed for Business; SCIM through managed users in an enterprise account |
| SOC 2 / ISO 27001 | SOC 2 Type II and ISO 27001 cover Copilot Business and Enterprise |
| Editors and surfaces | VS Code, Visual Studio, JetBrains, Xcode, Eclipse, Neovim, Zed, CLI, github.com |
Read from the vendor's public pages on Oct 16, 2026. No hands-on run is published yet. How we check
We have not published a sign-up run for GitHub Copilot yet. The access facts above were read from its sign-up and pricing pages on Oct 16, 2026.
We have not published API calls for GitHub Copilot yet.
GitHub Copilot lists 5 plans. The cheapest paid plan, Pro (individual), costs $10/mo.
Prices as listed, USD, excluding tax; observed 2026-10-16.
Source: github.com pricing page. Prices change; the observed date above is when our agent read them.
We read the pricing, docs, trust and legal pages of three AI coding assistants on the same day and scored them on the same six indicators. What 5 developers cost, how usage is counted, what the terms say about your code, and who should pick which.
