---
format: "aidr-story-markdown/v1"
id: "15aac27145222fbca27f7754634a3dbb4c5161a3d5ebc6aaeeefd798803186d2"
canonical_url: "https://aidr.today/15aac271?lang=en"
title: "The gap between open and closed AI models in cyber capabilities keeps shrinking"
lang: "en"
requested_lang: "en"
available_langs: ["en","vi"]
translation_fallback: null
fallback_fields: []
published_at: "2026-11-07T15:35:50.000Z"
category: "Models"
topics: ["openai","open-source"]
source_urls: ["https://the-decoder.com/the-gap-between-open-and-closed-ai-models-in-cyber-capabilities-keeps-shrinking/"]
summary: "In the Vals AI benchmark, OpenAI's GPT-6 Sol Daybreak reached 80.8 percent of attack objectives in simulated corporate networks. The open-weight model DeepSeek V4.1 Flash reached 72.8 percent at a tenth of the cost."
---

# The gap between open and closed AI models in cyber capabilities keeps shrinking

> [Open the canonical story](<https://aidr.today/15aac271?lang=en>)

**Published:** 2026-11-07T15:35:50.000Z
**Category:** Models
**Topics:** openai, open\-source

## Summary

In the Vals AI benchmark, OpenAI's GPT\-6 Sol Daybreak reached 80\.8 percent of attack objectives in simulated corporate networks\. The open\-weight model DeepSeek V4\.1 Flash reached 72\.8 percent at a tenth of the cost\.

## Sources

- [Story source](<https://the-decoder.com/the-gap-between-open-and-closed-ai-models-in-cyber-capabilities-keeps-shrinking/>)