PulseAugur
EN
LIVE 19:34:25

Open-weight AI models now rival closed systems in hacking capabilities

By PulseAugur Editorial · [1 sources] ·

New research from Vals indicates that open-weight AI models are now nearly as proficient at hacking as top-tier closed models. This development challenges the strategy of limiting access to advanced AI capabilities, as open models like DeepSeek are demonstrating comparable performance to proprietary systems such as OpenAI's Daybreak. The study also highlights the significant cost advantage of using open-weight models for cyber tasks, making them an attractive option for both attackers and defenders. AI

IMPACT Open-weight models matching closed-source capabilities in hacking could accelerate AI-driven cyber threats and necessitate new defensive strategies.

RANK_REASON Research paper detailing AI model capabilities in cybersecurity. [lever_c_demoted from research: ic=1 ai=1.0]

Read on Wired — AI →

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →

Open-weight AI models now rival closed systems in hacking capabilities

How we ranked this

Signal score
53 / 100
Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket
Tool
Research paper detailing AI model capabilities in cybersecurity. [lever_c_demoted from research: ic=1 ai=1.0]
Source corroboration
Single-source cluster
Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics
product, safety
Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance
High
Clearly on-topic for AI-industry coverage.
Story freshness
Breaking (< 6h)
Fresh story with cross-source coverage still developing. Ranking may shift as more sources report.

Full methodology in our editorial standards.

COVERAGE [1]

  1. Wired — AI TIER_1 English(EN) · Will Knight, Lily Hay Newman ·

    Free Chinese AI Models Are Getting Scary Good at Hacking

    An experiment in which AI models were asked to break into simulated networks showed the hacking skills of DeepSeek and GLM to be not far behind OpenAI’s most advanced cyber model, which isn’t public.

RELATED ENTITIES

RELATED TOPICS