Devin review
Devin (Cognition): cloud computer spec, pricing in USD, channels and published performance.
Data updated: 2026-11-05 09:38Cognition's AI software engineer: writes code, runs tests, and submits PRs on its own cloud machine; teams assign tasks via Slack, tickets, or web.
Classified as Work Agent (A digital employee for a team or organization). Cognition positions Devin as an AI software engineer within a team: each task is executed on a cloud-based Devin machine, assigned via Slack/tickets, billed per team seat, and categorized as a Work Agent. See definitions
Identity stability
Whether this cloud computer looks like the same person to websites over time.
| Metric | Finding (by source) |
|---|---|
| Egress IP: static / dynamic | Official shared static ranges Official docs list 7 static IPs plus one /26 range (AWS/Azure datacenter), shared by all customers; Cognition intends to keep them static. source |
| IP type: residential / ISP / business / datacenter | Official Datacenter |
| ASN stable | No data |
| Country / city stable | No data |
| Browser profile / cookies persist | Official Persist (after you save the browser profile) Official: log in once and have Devin save the browser profile (cookies/localStorage); every new session restores it. Profiles are organization-wide, so anyone in the org can start a session with those logins. source |
| Bring-your-own proxy / fixed egress | No data |
| Fit for long-term logins | OK for ordinary logins Not for sensitive accounts How it is judged → |
Does it have a VPS (cloud computer)
- VM (sandbox): A temporary execution environment provided by the platform; its lifecycle and state may not be stable. Everyday chat and simple tasks run in a VM; every agent, chat assistants included, has one, so it is not listed.
- VPS (cloud computer): A long-lived environment whose state persists, which the user or agent can enter again and again. A VPS always includes a VM, but not the other way round; we only track whether a product has a VPS.
| Metric | Finding (by source) |
|---|---|
| VPS | Official No Official: every session boots from a fresh copy of the organization's active snapshot, and changes made in a session are not written back; the saved browser profile is restored during setup. source |
Overview & spec
- Cloud computer (entry)
- -
- Vendor
- Cognition
- Type
- Work Agent
- Status
- Live; starting 2026-09-29, can log in with ChatGPT account to deduct OpenAI model usage.
- Launched
- 2024-12-10
- Model / framework
- SWE-2 (self-developed) + optional OpenAI/Claude/Gemini, etc.
- Channels
- Web, Devin Desktop, CLI, Slack
- Availability
- Global (enterprise can opt for dedicated deployment)
Plans Price: verifiable on official page
Official website price; Pro tier and above include cloud agent (Devin Cloud); overage purchased at API price. When logging in with ChatGPT Plus/Pro account, OpenAI model usage is deducted from ChatGPT quota. Official page last checked: 2026-11-05. USD shown; CNY converted at ¥6.68/USD.
| Plan | Price | Cloud computer | Notes |
|---|---|---|---|
| Free | Free | Same as entry tier | Limited quota, does not include Devin Cloud. |
| Pro | $20/mo | Same as entry tier | Includes Devin Cloud agent |
| Max | $200/mo | Same as entry tier | Quota significantly increased. |
| Teams | $80/mo | Same as entry tier | Team base price, plus $40/month per official seat, maximum 200 people. |
| Enterprise | Custom | Same as entry tier | Customized, billed by ACU, supports dedicated deployment. |
Apps & services
No verifiable list of connected apps/services yet (not published by the vendor or not collected).
Environment & limits
- Runtime
- Devin Cloud: one cloud-based Devin machine per session (configuration undisclosed); Enterprise version can use Outposts to execute commands on self-owned Devbox.
- Usage limits
- Specific figures for each tier's quota are not disclosed.
- Privacy / data
- Enterprise version offers controls like IP access control and dedicated deployment.
- Independently writes code, runs tests, and submits PRs.
- Slack and ticket assignment
- Devin Desktop (IDE) and CLI
- Sign in with ChatGPT to deduct OpenAI usage
- Enterprise Outposts / Dedicated deployment
Usage & published performance
No citable public performance data yet, and TANDIU has not run independent benchmarks.