【Source: InsiderInsiders Say: Openai'S New Project Is Behind Schedule, With The Compute Cost For Training The Model Reaching As High As $500 Million.】 December 23. Recently, Openai Launched The Next-Generation Model O3, Which Surpasses O1 In Software Engineering, Coding, Competitive Mathematics, And Even In Mastering Human Doctoral-Level Knowledge Of Natural Science. However, According To The Wall Street Journal, Openai'S New Artificial Intelligence Project Is Behind Schedule And Has Incurred Enormous Costs. It Is Currently Unclear When Or Even If It Will Succeed, As There Seems To Be Insufficient Data In The World To Make It Sufficiently Intelligent.It Is Reported That The Project'S Formal Name Is Gpt-5, Codenamed Orion, And Has Been In Progress For Over 18 Months. Sources Revealed That Openai'S Closest Partner And Largest Investor, Microsoft, Originally Expected To See The New Model Around Mid-2024. Openai Has Also Conducted At Least Two Large-Scale Training Sessions, Taking Months To Process Massive Amounts Of Data To Make The New Model "Orion" More Intelligent.However, According To Sources Close To The Project, Each Training Session Has Encountered New Issues, And The Software'S Performance Has Not Met Researchers' Expectations. Even In The Best-Case Scenario, Orion'S Performance Is Improved Compared To Openai'S Current Model, But It Is Not Yet Sufficient To Justify The Massive Costs Required To Sustain The Operation Of The New Model. Based On Public Data And Private Estimates, The Computational Cost Of A Single Six-Month Training Run Could Be As High As Around $500 Million