Part of the What will be the state of AGI by the end of 2030? topic →
This market resolves YES if no system is broadly recognized as matching or exceeding average human cognitive abilities across the full range of tasks, including novel situations, by the end of 2030. Current systems show rapid gains in math, coding and science, but the International AI Safety Report describes their capabilities as jagged, and METR’s assessment of GPT-5.6 Sol found highly uncertain task-horizon estimates and no evidence that it enabled fully automated AI research and development. A separate low-volume external forecast puts AI achieving AGI by 2030 near an even chance, but it does not clearly match this market’s demanding definition; expert scenarios also leave substantial uncertainty. These measurement and generalization gaps keep non-arrival the likeliest single outcome, while continued progress leaves a substantial chance of AGI by the deadline.
A review found expert and superforecaster predictions lagged observed results on several near-term benchmarks and company revenue, while emphasizing that the evidence remains insufficient to judge broad societal-impact forecasts.
After employee and researcher warnings about rapidly advancing capabilities and containment concerns, leaders including Anthropic’s Dario Amodei and OpenAI’s Sam Altman publicly supported pacing frontier development; reporting also described coordination and competition obstacles.
OpenAI launched Astra on September 3, presenting advances in computer use, software, cybersecurity and science; President Greg Brockman said he personally thought “we’re there.” This is a high-profile company claim, not a consensus finding under the question’s full-range, novel-situation definition.
A Nature perspective distinguishes autonomy and efficacy from generality, helping separate agentic benchmark progress from the market’s AGI definition.
Recent expert panels favored moderate rather than rapid progress by 2030, while forecasting conditional health and economic gains under rapid progress alongside non-negligible catastrophe and harm-event risks.
Will AGI be developed and be net positive to the world by the end of 2030?
Will AGI be developed and be net negative to the world by the end of 2030?
Will superintelligent software attack people at will by 2036?
Will recursive self-improvement be observed in AI by 2030?
Will super intelligent software extort people at will to gain advantage within the next 5 years?
Will any meaningful federal AI regulation be enacted into law by 2029?
