GILD verdict
Choose Devin when a team needs shared billing, admin control over data and access, or repositories beyond GitHub.
Choose Jules when individual developers want to delegate scoped GitHub work at no cost or on a Google AI plan's higher task limits. On public GitHub, PR Arena counts 84.0% of Jules's ready pull requests as merged and 64.6% of Devin's (October 28, 2026). Both return pull requests a human should approve.
Devin is the team purchase and Jules the individual one. Cognition sells Devin with Teams seats, admin-set training opt-outs, several Git hosts, and Slack and Linear triggers. Google built Jules around individual developers on GitHub, with a free tier and Google AI plan limits. In each, a task goes in, the agent works in its own cloud machine, and a pull request comes back.
Choosing between Jules and Devin
Choose Jules if
- Individual developers want to evaluate delegated work at no cost on GitHub repositories.
- A fixed number of tasks per day suits you better than metered usage and credits.
- You want scheduled maintenance tasks and automatic CI fixes on the agent's own pull requests.
Choose Devin if
- You need one subscription for a team, with full and flex seats and shared credits.
- Repositories live on GitLab, Bitbucket, or Azure DevOps as well as GitHub.
- Work should start from Slack, Linear, or Jira, with an admin controlling training opt-out and repository access.
Side by side
| Dimension | Jules | Devin |
|---|---|---|
| Individual priceList prices in US dollars. Jules counts tasks, Devin a daily and weekly quota plus credits, so included work is not directly comparable. | Free plan; Google AI Pro at US$19.99 per month | Free plan; Pro at US$20 per month; Max at US$200 per month |
| Team plan | None; paid plans are for personal @gmail.com accounts only | Teams: US$40 per full seat per month, US$80 monthly minimum, up to 200 members |
| Beyond the plan | New tasks are blocked until the rolling 24-hour window resets | Pay-as-you-go on-demand credits past the quota |
| Repository hosts | GitHub only | GitHub, GitLab, Bitbucket, and Azure DevOps |
| Where tasks start | Web app, CLI, alpha REST API, GitHub Action, and GitHub issues labeled jules | Web app, Slack, Linear, Jira, /devin pull request comments, and the API |
| Training default | Google states Jules does not train on private repository content | May train on your data by default; paid plans can opt out, and on Teams only an admin can |
| Repository access | Each user installs the Google Labs Jules GitHub app on chosen repositories | A GitHub organization admin connects all or selected repositories for the Devin organization |
| Merged pull requests on public GitHubPR Arena's count of pull requests each agent opened on public GitHub, updated October 28, 2026. Ready means the pull request left draft. | 84.0% of 245,534 ready pull requests12 | 64.6% of 311,181 ready pull requests12 |
What practitioners report, side by side
Each product's figures come from its own practitioners, for the group you pick.
Overall signal
All practitioners rate Devin +37, 37 points above Jules at 0. Most of the gap is in praise, with 30 of Devin's 53 sources positive against 26 of Jules's 68. Devin has 12 negative and Jules 24.
The score counts the authors from all practitioners that take a side, one vote each with their latest view, and is how far positive votes lead negative ones, per 100 votes, so Jules's 0 rests on votes split 24 positive, 18 mixed and 24 negative and Devin's +37 on 30, 10 and 11.
A lead of a fifth of the votes reads positive and a lead of half strong positive, with the same lines below zero. The dots show each source, positive, mixed, negative. Use a task pilot to measure accepted work and review time.
Footprint on the Agentic Software Development stack
They differ on two of 12 layers. On Code review, Jules touches it and Devin supports it. On Application security, Jules touches it and Devin supports it. Coverage is each product's placement on this stack, as its product page shows it. Core is a job the product owns; Supports and Touches are jobs it helps with. Coverage says nothing about how well a product does them.
| Layer | Jules | Devin |
|---|---|---|
| Validate | ||
| Code review | Touches | Supports |
| Application security | Touches | Supports |
The 10 layers where they match
| Layer | Jules | Devin |
|---|---|---|
| Plan | ||
| Work planning | Supports | Supports |
| Build | ||
| Codebase context | Supports | Supports |
| Coding | Core | Core |
| Agent runtime | Supports | Supports |
| Tool execution | Core | Core |
| Validate | ||
| Debugging | Supports | Supports |
| Testing | Supports | Supports |
| Ship | ||
| Source control | Touches | Touches |
| CI/CD | Touches | Touches |
| Operate | ||
| Observability | Touches | Touches |
What practitioners praise
Among all practitioners, Free tier & plan value and Maintenance & chores reach 5 sources from 3 authors only for Jules, at 11 of 12 and 6 of 6. Devin needs a pilot on the same repository tasks before you choose on either. Of the three themes both carry, End-to-end execution differs most. It is praised in all 17 Jules sources that discuss it, against 17 of 18 for Devin. Codebase understanding and Migrations & repetitive changes reach 5 sources from 3 authors only for Devin, at 14 of 16 and 7 of 7. Jules needs a pilot on the same repository tasks before you choose on either.
| Praised for | Jules | Devin |
|---|---|---|
| End-to-end execution | 17 of 17 | 17 of 18 |
| Async delegation | 14 of 14 | 16 of 16 |
| Developer experience | 5 of 10 | 8 of 15 |
| Free tier & plan value | 11 of 12 | Not a theme in its sources |
| Maintenance & chores | 6 of 6 | Not a theme in its sources |
| Codebase understanding | Not a theme in its sources | 14 of 16 |
| Migrations & repetitive changes | Not a theme in its sources | 7 of 7 |
Each count is the sources that praise a theme, or raise it as a concern, out of those that discuss it. A row pairs the two products when both carry a theme under the same or an equivalent name.
A count ending in "so far" has under 5 sources from 3 authors among all practitioners. "Not a theme in its sources" means GILD found no theme of that name in that product's sources.
What practitioners raise as concerns
Devin's sources raise Needs Close Steering more often than Jules's. It is raised in 3 of the 7 Jules sources that discuss it, against 9 of 10 for Devin. Jules alone has four concern themes that reach 5 sources from 3 authors. Unreliable task completion and GitHub workflow friction lead, at 24 of 25 and 10 of 12. Devin alone has four concern themes that reach 5 sources from 3 authors. Unpredictable task success and Usage and cost lead, at 11 of 17 and 11 of 18.
| Raised as a concern | Jules | Devin |
|---|---|---|
| Requires strong review | 19 of 20 | 11 of 15 |
| Needs Close Steering | 3 of 7 | 9 of 10 |
| Unreliable task completion | 24 of 25 | Not a theme in its sources |
| GitHub workflow friction | 10 of 12 | Not a theme in its sources |
| Task speed | 8 of 9 | Not a theme in its sources |
| VM environment setup | 5 of 5 | Not a theme in its sources |
| Unpredictable task success | Not a theme in its sources | 11 of 17 |
| Usage and cost | Not a theme in its sources | 11 of 18 |
| Environment & context setup | Not a theme in its sources | 4 of 8 |
| Session speed | Not a theme in its sources | 3 of 5 |
12 sources
- Jules docs: Limits and plans
- Google One: Google AI plans and Jules limits
- Jules docs: FAQ
- Jules docs: Managing tasks and repos
- Jules docs: Running tasks
- Jules changelog: CI Fixer and commit authoring (Feb 19, 2026)
- Devin docs: Self-serve plans
- Devin docs: Security at Cognition
- Devin docs: GitHub integration
- Devin docs: Integrations overview
- Devin docs: Environment setup
- PR Arena: AI coding agent pull request leaderboard (updated Oct 28, 2026 UTC)