Skip to main content

AI Launch Profile

Cognition launches Devin 2.2 with computer use, self-verification, and autofix

Cognition released Devin 2.2 with desktop computer use, end-to-end testing, self-verification, review autofix, faster startup, and a redesigned interface.

Editorial image for Cognition launches Devin 2.2 with computer use, self-verification, and autofix AI launch

At a glance

Launch Snapshot

Company
Cognition
Launch date
February 24, 2026
Launch type
Major Update
Category
AI Agents, AI Coding Tools, AI Developer Tools
Audience
Engineering teams that want a coding agent to plan, implement, test, record evidence, fix issues, and prepare pull requests.
Pricing
Commercial Devin credits/pricing; new users received starter credits at launch according to Cognition.
Free plan
Yes
API
No
Open weights/source
No

Verification & Sources

Evidence state
Recheck due
Source links
2
Freshness
Needs recheck: checked June 8, 2026
Last updated
June 8, 2026
What this evidence state means
Definition
The claim was previously checked, but its review window expired or a material change may have invalidated it.
Required provenance
The prior evidence and check date are retained, together with the expiry or change signal that triggered recheck.
Owner
Kingy freshness queue owner and assigned editorial reviewer
Freshness rule
This is already outside its freshness rule. It must not be presented as current until reviewed against current evidence.
Disputes and corrections
Use “Suggest a correction” on the record. Kingy editorial reviews the cited evidence, records material corrections, and changes or removes the state when it is not supported.

Key source checks

Suggest a correction

Form submissions, correction notes, score details, URLs, and analytics events may be stored for editorial review, spam prevention, product improvement, and follow-up. Do not submit secrets, unreleased financials, private customer data, or regulated personal data through these forms.

Kingy AI Take

This was a major Devin update because it tightened the full loop from code generation to computer-use testing and autofix.

Who it is for

Engineering teams that want a coding agent to plan, implement, test, record evidence, fix issues, and prepare pull requests.

What feels promising

Computer-use verification and screen recordings make agent work easier to review before merging.

What feels unproven

Autonomous coding still needs careful scoping, test quality, secrets controls, and human review for production changes.

Editorial submissions and sponsor-fit reviews are separate. Payment does not influence Kingy scores, verdicts, rankings, evidence labels, or publication decisions.