Now Reading: OpenAI launches GPT-5.6-Cyber model for cyber defence
-
01
OpenAI launches GPT-5.6-Cyber model for cyber defence
OpenAI has launched GPT-5.6-Cyber, a specialised model for cybersecurity, available through its Daybreak Red access programme. According to Gadgets360, the company says an internal test showed the model outperformed its general-purpose AI…
The Story in Brief
OpenAI has launched GPT-5.6-Cyber, a specialised model for cybersecurity, available through its Daybreak Red access programme. According to Gadgets360, the company says an internal test showed the model outperformed its general-purpose AI on defence tasks and gave fewer refusals. Early-access partners include SpecterOps, SentinelOne and Palo Alto Networks. They used the model to improve defensive workflows. OpenAI says trusted customers received access first.
The Indian Opinion
The hype around AI-powered cybersecurity often obscures real limits. Gadgets360 reports only OpenAI's internal test showed gains; we lack independent verification. Past AI models promised much but failed against novel threats. The early partners are security vendors themselves, not neutral assessors. Until a respected third party like a national cyber agency publishes a public benchmark, this launch remains a marketing event. Will the model handle zero-day exploits or just known patterns?
Source: gadgets360.com
This story was synthesised by AI from the source linked above.
