After months of speculation, OpenAI’s Chief Technology Officer, Mira Murati, finally shed some light on the capabilities of the much-anticipated GPT-5 (or whatever its final name will be).
In an interview with Dartmouth Engineering, Murati describes the jump from GPT-4 to GPT-5 as a significant leap in intelligence. She compares GPT-3 to toddler-level intelligence, GPT-4 to smart high-schooler intelligence, and GPT-5 to achieving a "Ph.D. intelligence for specific tasks."
Release Not Until Late 2025 or Early 2026
Interestingly, the interviewer pressed Murati for a specific release date. While there were earlier rumours of a late 2023 launch and more recent reports suggesting a summer 2024 release, Murati clarified that GPT-5 is still a year and a half away, potentially pushing its release to late 2025 or early 2026.
Confirmation of Ph.D.-Level Intelligence for Specific Tasks
Murati’s remarks echo earlier predictions about GPT-5’s potential. Similar to Microsoft CTO Kevin Scott’s comments about next-gen AI systems passing Ph.D. exams, Murati highlights GPT-5’s advanced memory and reasoning capabilities.
However, Murati clarifies that this "Ph.D.-level" intelligence is task-specific. While these systems can achieve human-level performance in certain tasks, they still lag behind in many others.
SEE ALSO: New AI Scan From Oxford Promises Early Detection And Improved Heart Attack Prevention
{Categories} _Category: Platforms{/Categories}
{URL}https://in.mashable.com/tech/77593/openais-gpt-5-pushed-back-to-late-2025-but-promises-phd-level-abilities{/URL}
{Author}unknown{/Author}
{Image}https://sm.mashable.com/t/mashable_in/photo/default/7_9zpf.1200.jpg{/Image}
{Keywords}{/Keywords}
{Source}Platforms{/Source}
{Thumb}{/Thumb}