# Exclusive: Cogent Security's VR-1.1 model outscores Claude Code and Codex on vulnerability repair

DevFeed: [Exclusive: Cogent Security's VR-1.1 model outscores Claude Code and Codex on vulnerability repair](<https://devfeed.tech/articles/exclusive-cogent-security-s-vr-1-1-model-outscores-claude-code-and-codex-on-vulnerability-repair-114238.md>)

Original publisher: [Read original article](<https://siliconangle.com/2026/11/03/exclusive-cogent-securitys-vr-1-1-model-outscores-claude-code-and-codex-on-vulnerability-repair/>)

Author: Duncan Riley

Published: 2026-11-03T14:00:00Z

Content type: news

Language: en

Sources: [SiliconANGLE](<https://devfeed.tech/sources/siliconangle.md>)

Topics: [vulnerability management](<https://devfeed.tech/topics/vulnerability-management.md>), [patches](<https://devfeed.tech/topics/patches.md>), [code security](<https://devfeed.tech/topics/code-security.md>), [Claude Code](<https://devfeed.tech/topics/claude-code.md>), [codex](<https://devfeed.tech/topics/codex.md>), [Security](<https://devfeed.tech/topics/security.md>), [ai code review](<https://devfeed.tech/topics/ai-code-review.md>), [Code quality](<https://devfeed.tech/topics/code-quality.md>), [Claude](<https://devfeed.tech/topics/claude.md>), [OpenAI](<https://devfeed.tech/topics/openai.md>), [Test-driven development](<https://devfeed.tech/topics/tdd.md>)

Tags: [ai](<https://devfeed.tech/tags/ai.md>), [ai-agents](<https://devfeed.tech/tags/ai-agents.md>), [ai-benchmarks](<https://devfeed.tech/tags/ai-benchmarks.md>), [anthropic](<https://devfeed.tech/tags/anthropic.md>), [anthropic-pbc](<https://devfeed.tech/tags/anthropic-pbc.md>), [bain-capital-ventures](<https://devfeed.tech/tags/bain-capital-ventures.md>), [benchmark](<https://devfeed.tech/tags/benchmark.md>), [claude](<https://devfeed.tech/tags/claude.md>), [claude-code](<https://devfeed.tech/tags/claude-code.md>), [claude-mythos-preview](<https://devfeed.tech/tags/claude-mythos-preview.md>), [code](<https://devfeed.tech/tags/code.md>), [codex](<https://devfeed.tech/tags/codex.md>), [cogent](<https://devfeed.tech/tags/cogent.md>), [cogent-security](<https://devfeed.tech/tags/cogent-security.md>), [cybersecurity](<https://devfeed.tech/tags/cybersecurity.md>), [exclusive](<https://devfeed.tech/tags/exclusive.md>), [frontier](<https://devfeed.tech/tags/frontier.md>), [frontier-models](<https://devfeed.tech/tags/frontier-models.md>), [geng-sng](<https://devfeed.tech/tags/geng-sng.md>), [greylock-partners](<https://devfeed.tech/tags/greylock-partners.md>), [harbin-institute-of-technology](<https://devfeed.tech/tags/harbin-institute-of-technology.md>), [inc](<https://devfeed.tech/tags/inc.md>), [kuaishou-technology](<https://devfeed.tech/tags/kuaishou-technology.md>), [livecvebench](<https://devfeed.tech/tags/livecvebench.md>), [management](<https://devfeed.tech/tags/management.md>), [model](<https://devfeed.tech/tags/model.md>), [news](<https://devfeed.tech/tags/news.md>), [openai](<https://devfeed.tech/tags/openai.md>), [patching](<https://devfeed.tech/tags/patching.md>), [repair](<https://devfeed.tech/tags/repair.md>), [security](<https://devfeed.tech/tags/security.md>), [the-latest](<https://devfeed.tech/tags/the-latest.md>), [vineet-edupuganti](<https://devfeed.tech/tags/vineet-edupuganti.md>), [vr-1](<https://devfeed.tech/tags/vr-1.md>), [vr-1-1](<https://devfeed.tech/tags/vr-1-1.md>), [vulnerability](<https://devfeed.tech/tags/vulnerability.md>), [vulnerability-management](<https://devfeed.tech/tags/vulnerability-management.md>), [vulnerability-remediation](<https://devfeed.tech/tags/vulnerability-remediation.md>)

## AI overview

On LiveCVEBench's 190 repair tasks, Cogent Security says VR-1.1 produced working fixes for 57.9%, compared with 45.3% for Claude Code and Codex. Each task places an agent in a real development environment with a vulnerability report and affected code; a fix must pass separate security and functionality tests. Cogent says VR-1.1 can first reproduce a flaw with a proof of concept, trace it through related code paths, make a small fix, and submit the patch to a fresh-context review that checks for bypasses and newly introduced bugs. It also generates regression tests and carries failure details into subsequent attempts. Cogent blocked outbound network access during repairs, while inference budgets differed among the systems. The excerpt does not provide those budgets or further benchmark methodology.

## Source excerpt

Vulnerability management startup Cogent Security Inc. today released VR-1.1, a new version of its frontier reasoning model that writes fixes for newly disclosed software vulnerabilities. On a public repair benchmark, the model outscored Anthropic PBC's Claude Code and OpenAI Group PBC's Codex. The release goes after the part of vulnerability management that has not sped [...] The post Exclusive: Cogent Security's VR-1.1 model outscores Claude Code and Codex on vulnerability repair appeared first on SiliconANGLE.