Corridor is joining Cognition to work on Devin Security
Corridor's co-founders said the security team is joining Cognition. The posts do not give deal terms or a date for a product change.
Every Devin update in one feed — release notes, features and changes, straight from the source.
22 updates
Corridor's co-founders said the security team is joining Cognition. The posts do not give deal terms or a date for a product change.
Devin Security 2.0 adds scans with Claude Mythos and other frontier models, threat-model prioritization, and parallel remediation sessions that open PRs.
The refreshed Devin Desktop app centres on an agent command center that Cognition says is smoother and more polished than before.
The open-source agent plugin helps secure systems against recent cybersecurity threats. It works with any agent and in Devin today.
The open-source Vulnerability Fixes plugin finds affected PHP and Node repos, tests before and after patching, and opens one pull request per repo.
Blanket usage caps hold back your most efficient engineers.
Cognition's on-call agent investigates alerts and bug reports, and scored 100% end-to-end on the SREGym incident benchmark.
Support teams can trigger Devin from Pylon tickets, and new Devin customers on Pylon get $15,000 in credits.
Available in Devin Desktop and CLI at 75% off list through October 27, 2026, with a 60.5% score on Devin's FrontierCode 1.1 benchmark.
Haiku 5.5 scores 58.4% on Devin's FrontierCode 1.1 and can serve as the Fusion sidekick, which you can try now in the Devin CLI with Opus 5.5 leading.
Memories are personal to each user in an organization, kept in a Git-backed drive, and a daily Dreaming session consolidates them and removes stale notes.
GPT-6.1 Sol matches GPT-6 Sol on Devin's FrontierCode 1.1 benchmark at 44% to 57% lower cost per task, in Devin Desktop and CLI.
Devin Pro, Max and Teams users can bill OpenAI model usage in Cloud, Desktop and CLI to an eligible ChatGPT Plus or Pro plan.
The integration is available now to MongoDB and Cognition customers, and early joint testing cut some migration work from five or six hours to about one.
Devin cites a 64.4% FrontierCode 1.1 score for the model, plus Anthropic's figures of 30% faster output and up to 30% lower cost per task.
Fusion and Normal mode cost 30-40% less, Ultra 15-20% less and Devin Review up to 70% less, with Fusion scoring 68.8 on FrontierCode 1.1.
Available to all Devin customers, with six Microsoft 365 connections that use delegated per-user permissions and read-only minimums.
Both OpenAI models are live in Devin Desktop and CLI, with Luna coming in under $0.10 per task on Devin's FrontierCode 1.1 benchmark.
Claude Opus 5.5 tops Devin's FrontierCode 1.1 benchmark at under a tenth of Fable 5's cost per task, in Devin Desktop and CLI.
Devin's benchmark finds the model strongest on hard Java, Go and Ruby backend tasks, but it scores below Grok 4.6 overall because it over-scopes.
Devin CLI gains cloud handoff, session resume and SSH access to Devin's VM, with free SWE-2 sessions offered until October 8.
Devin's new /scan command investigates a codebase against a stated goal, ranks the findings and opens pull requests for the ones you choose.