Unverified50% confidenceBenchmarkExact time
By July of this year, combined with UltraFast, the model was already able to complete more than one-third of day-long research tasks without human intervention.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
9/30/2026
First Seen
Valid until: 12/29/2026
Sources
Related Entities
Related Claims
UnverifiedStage one focuses on foundational concepts, including large language models, how prompts work, and the basic principles of AI agents, with a duration of approximately 74 minutes.60% similarUnverifiedWith an AI-assisted performance testing approach, performance testing work that originally took 30 days can now essentially be completed in 1 day.58% similarUnverifiedPilot experiments had three coding agents and eight models each run eight 20-minute single-pass tasks, with label-aware F1 showing extremely high variance ranging from 0.143 to 0.892.58% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/960904API
curl https://kongchang.com/api/v1/knowledge/claims/960904MCP
get_claim(id=960904)