OpenAI expects its AI systems to reach the capability of a research intern by September 2026 and to function as a fully autonomous AI researcher by March 2028, according to internal targets attributed to Chief Scientist Jakub Pachocki.
Internal Roadmap · OpenAI
Building a "Research Lab in a Data Center"
OpenAI has set internal targets to reach research-intern capability by September 2026 and a fully autonomous AI researcher by March 2028 — automating the scientific process that drives AI development itself.
Sep 2026
Target for an AI research intern — autonomous work spanning several days
Mar 2028
Target for a fully automated researcher — likely a multi-agent system
$1.4T
Planned compute investment underpinning the ambition
The task horizon: how long AI works unattended
The practical benchmark separating an assistant from an independent researcher is the length of time it can run autonomously.
Minutes–hours
Today's assistants
Several days
Intern-level target
The feedback loop OpenAI is chasing
AI that helps design and improve the next generation of AI — accelerating discovery itself.
AI runs experiments
→
Proposes follow-up work
→
Improves next AI
Optimists
Today's models already approach intern-level capability on narrow tasks, and automated research is edging closer.
Skeptics
Benchmark scores don't capture the messy, open-ended nature of real science — the timeline may be too aggressive.
These remain internal goals, not shipped products . No independent, real-world evaluation of intern-level research automation has been published — the September 2026 milestone will be the first concrete test.
Continue reading The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.
Already purchased? Sign in ✓ Signed in — this article isn’t included in your current plan.Unlocking the full article…