Skip to content
Hot eventDeveloping

Vals study finds open-weight DeepSeek nearly matches OpenAI's gated Daybreak on network hacking tasks

1 reports1 reporting sourcesUpdated 42 min ago

Start here

AI digest

Vals tested AI models on multi-step network intrusion tasks against mock corporate networks and found that open-weight models now nearly match the best closed ones. OpenAI's Daybreak, which is available only to trusted partners, completed 81 percent of assigned tasks on average, while the open-weight DeepSeek completed 73 percent; the two scored equally on some tasks. According to Vals, an attack costs $17 to $46 in tokens with GPT-6-Sol, versus $2 to $3 with DeepSeek. Researcher Nikil Ravi said gated access at frontier labs may no longer buy much time before such capabilities spread.

Generated from reports · updated 34 min ago

Report timeline

Follow the reports to see the event from different sides.

November 16
  1. Hacker News · AI
    Vals study finds open-weight DeepSeek nearly matches OpenAI's gated Daybreak on network hacking tasks

    Vals tested how well AI models could carry out multi-step network intrusion tasks on mock corporate networks and found open-weight models now nearly match the best closed ones. OpenAI's Daybreak, available only to trusted partners, completed 81 percent of assigned tasks on average, while the open-weight DeepSeek completed 73 percent, with the two scoring equally on some tasks. Vals says an attack costs $17 to $46 in tokens for GPT-6-Sol versus $2 to $3 for DeepSeek, and researcher Nikil Ravi says gated access at frontier labs may no longer buy much time before such capabilities spread.

Heat for this event

Not enough continuous observations yet to draw a trend.