When will OpenAI achieve AGI?
Short Answer
Both the model and the market expect OpenAI to achieve AGI before 2030, with no compelling evidence of mispricing.
1. Executive Verdict
- Executive statements and internal roadmaps indicate an accelerated AGI development outlook.
- Empirical evidence suggests accelerated AGI timelines, projecting rapid benchmark saturation.
- US federal AI safety reviews may significantly alter development timelines before 2030.
- Cautious AGI forecasts often cite significant resource and technical challenges.
- OpenAI emphasizes scaling and autonomous agentic workflows for achieving AGI.
- METR's 'Task Standard' framework is likely to validate AGI claims before 2030.
Who Wins and Why
| Outcome | Market | Model | Why |
|---|---|---|---|
| Before 2027 | 8.2% | 8.6% | Executive and competitive expectations point to an earlier AGI arrival window. |
| Before 2028 | 29.5% | 28.8% | Executive projections and OpenAI's roadmap target AGI-significant capabilities within this timeframe. |
| Before 2030 | 45.0% | 39.1% | Executive statements and internal roadmaps suggest an accelerated outlook for AGI development. |
Current Context
Industry leaders project AGI timelines, though definitions remain debated. Artificial General Intelligence (AGI) is generally defined as an AI system capable of performing most economically valuable cognitive tasks at or above human levels, often with a specific focus on autonomous reasoning and scientific discovery [^][^]. As of June 2026, prominent figures in the AI industry have moved AGI timelines forward. OpenAI CEO Sam Altman and Anthropic CEO Dario Amodei have suggested potential AGI arrival windows between 2026 and 2028 [^][^]. DeepMind's Demis Hassabis, however, offers more cautious estimates, typically in the 2028-2030 range [^][^].
Prediction markets show caution regarding OpenAI's near-term AGI achievement. As of mid-2026, these markets assign low-to-moderate probabilities, often between 15-30%, for an AGI announcement by OpenAI before 2027 [^][^][^][^][^]. Confidence in AGI arrival increases significantly, with probabilities exceeding 50%, for achievement by 2030 [^][^][^][^][^].
OpenAI's strategy involves AI systems accelerating internal research. The company's internal roadmap, updated in June 2026, emphasizes the development of 'automated AI researchers' [^][^]. OpenAI's internal target is for AI systems to perform a significant fraction of research in conjunction with human researchers by March 2028 [^][^].
2. Market Behavior & Price Dynamics
Historical Price (Probability)
Outcome probability
Date
This market has demonstrated a sideways trading pattern, oscillating within a relatively tight range. The price for a 'YES' outcome has fluctuated between a low of 6.7% and a high of 10.0%, indicating these levels are acting as short-term support and resistance, respectively. The contract began trading at 7.0% and is currently priced at 8.2%, showing a slight upward drift but no definitive breakout. The overall price action suggests a market in equilibrium, with traders finding a balance point below the 10% probability mark. There have been no dramatic price spikes or drops, but rather a gradual climb from the low point established earlier in the period.
The recent upward movement from the 6.7% support level toward the current 8.2% price appears to coincide with recent context where industry leaders reportedly suggested potential AGI arrival windows between 2026 and 2028. This expert commentary may have bolstered confidence among 'YES' traders, preventing the price from falling further and encouraging a modest rally. The total traded volume of 7,066 contracts indicates consistent engagement and liquidity in the market. However, without daily volume data, it is difficult to determine if specific price moves were backed by high conviction or simply occurred on low volume.
Overall, the chart suggests that market sentiment remains cautiously skeptical but is responsive to expert forecasts. The low single-digit probability reflects a consensus that AGI, as defined by the market rules, is a long-shot event by the resolution date. The stable, range-bound trading implies that while traders acknowledge the optimistic timelines projected by some industry figures, they are not yet convinced enough to price in a significantly higher probability. The market appears to be in a "wait-and-see" mode, looking for more concrete evidence of a technological breakthrough before committing to a stronger directional trend.
3. Market Data
Contract Snapshot
The market resolves to YES if OpenAI announces they have attained AGI by December 31, 2029, with the outcome verified directly from OpenAI. If no such announcement is made by this deadline, the market resolves to NO, closing at 11:59 PM EST on December 31, 2029. Trading is prohibited for individuals employed by Source Agencies or possessing material, non-public information.
Available Contracts
Market options and current pricing
| Outcome bucket | Yes (price) | No (price) | Last trade probability |
|---|---|---|---|
| Before 2027 | $0.08 | $0.92 | 8% |
| Before 2028 | $0.30 | $0.71 | 30% |
| Before 2030 | $0.46 | $0.55 | 45% |
Market Discussion
Traders currently assign a 45% probability to OpenAI achieving AGI before 2030, though confidence in earlier dates like before 2027 is much lower at 8.2%, with all probabilities trending downwards. Arguments for an earlier achievement reference reports of a breakthrough project dubbed "Q*," capable of grade-school math and considered a significant step toward AGI. Conversely, arguments against near-term AGI highlight expert opinions from a congressional hearing, where AI specialists dismissed the likelihood of AGI within the next five years.
4. What empirical evidence from recent AI model performance supports the accelerated AGI timelines (pre-2028) suggested by executives at OpenAI and Anthropic?
| Cognitive benchmark saturation | 2028-2030 [^] |
|---|---|
| AI task completion growth | Doubling every three months (early 2026) [^][^] |
| Prediction market AGI probability | 13-25% by end of 2026 or before 2027 (mid-2026) [^][^][^][^] |
Empirical evidence supports accelerated AGI timelines, projecting rapid benchmark saturation. This evidence indicates a saturation of cognitive benchmarks by 2028-2030 [^]. As of early 2026, AI task completion time horizons are reportedly doubling every three months [^][^]. Additionally, simulation models suggest potential 'explosive' growth within six years, largely driven by automated AI research and development feedback loops [^].
Executives at OpenAI and Anthropic foresee powerful AI by 2028. These empirical findings align with public statements from leaders like Sam Altman and Dario Amodei, who have adjusted their expectations for AGI or 'powerful AI' systems to emerge within a 2026-2028 window [^][^][^][^][^]. Specifically, Anthropic has publicly documented expectations for 'powerful AI' systems to appear by late 2026 or early 2027 [^][^][^].
Prediction markets offer more conservative AGI timelines than executives. In contrast to executive predictions, markets such as Polymarket and Kalshi provide more cautious estimates. As of mid-2026, these platforms assign only a ~13-25% probability to OpenAI achieving or announcing AGI by the end of 2026 or before 2027 [^][^][^][^].
5. How do the AGI development roadmaps of OpenAI and competitors like Anthropic and Google DeepMind compare in their core technical approaches and stated timelines?
| OpenAI AGI Timeline | 2027-2029 [^][^][^][^] |
|---|---|
| Anthropic Powerful AI Timeline | 2026-2027 [^][^] |
| Google DeepMind Human-Level AI Timeline | 2028-2030 [^] |
Leading AI labs pursue AGI through distinct technical strategies and definitions. OpenAI emphasizes scaling and autonomous agentic workflows (Level 2-3), defining Artificial General Intelligence (AGI) by its ability to achieve economically valuable task superiority, such as reaching a $100 billion profit milestone [^][^]. Anthropic focuses on Constitutional AI and robust, safety-driven scaling, aiming to develop "powerful AI" systems that surpass human researchers in most domains [^][^]. Google DeepMind prioritizes scientific discovery, breakthroughs in reasoning, and mechanistic interpretability, conceptualizing AGI through its capacity for scientific invention and general problem-solving [^][^].
Stated timelines for AGI achievement are ambitious, varying slightly among key players. Anthropic anticipates achieving powerful AI systems by 2026-2027 [^][^]. OpenAI leadership indicates a clear path to AGI between 2027 and 2029 [^][^][^]. Google DeepMind offers a slightly more conservative forecast, projecting human-level AI between 2028 and 2030 [^]. As of June 2026, prediction markets reflect considerable uncertainty regarding these timelines, assigning approximately a 45-59% probability to OpenAI achieving AGI before 2030, with lower but still significant probabilities for an arrival before 2027 or 2028 [^][^][^].
6. What potential regulatory frameworks or international agreements on AI safety could materially alter OpenAI's development timeline before 2030?
| Federal Safety Review Order Date | June 2, 2026 [^] |
|---|---|
| Federal Review Period | up to 30 days before deployment [^] |
| OpenAI Projected Research AI | March 2028 [^][^] |
US federal AI safety reviews could significantly alter development timelines. A US Presidential executive order, signed on June 2, 2026, established a federal safety review process for frontier AI models. This mandatory process necessitates a review up to 30 days before deployment, which has the potential to impact the development schedules of companies such as OpenAI [^]. Notably, OpenAI itself is advocating for more rigorous, mandatory oversight than is currently in place under US policy, proposing evaluations by civilian-led agencies and supporting the creation of a global AI pause watchdog [^][^]. This federal review process could materially affect OpenAI's aspiration for AI to perform a significant portion of its own research by March 2028 [^][^].
International AI safety agreements offer amplified timeline impacts. While domestic policies evolve, international regulatory efforts are currently fragmented but are progressing towards formalizing risk management strategies [^][^]. The influence of regulation on AI development timelines is subject to a 'racing multiplier effect,' where globally coordinated agreements demonstrate 2-10 times greater effectiveness compared to unilateral slowdowns [^][^]. This suggests that such international frameworks have the potential to significantly modify the trajectory of AI development prior to 2030 [^][^].
7. Which emerging AGI evaluation frameworks, from groups like METR or academic institutions, are most likely to be used to validate an AGI claim before 2030?
| Primary AI agent evaluation standard | METR's 'Task Standard' [^][^][^][^] |
|---|---|
| Emerging AGI validation approach | 'Mosaic' approach combining benchmarks and standards [^][^] |
| AGI definition framework | 'Levels of AGI' ontology [^] |
Emerging frameworks aim to validate AGI claims before 2030 through specific evaluation methodologies. One key approach is METR's 'Task Standard,' which evaluates AI agents by measuring their autonomous capabilities across modular task families in domains such as AI R&D and cybersecurity [^][^][^][^]. Another notable method, the 'mosaic' approach, integrates standard benchmarks like GPQA, SWE-bench, and ARC-AGI with auditable, multi-dimensional scorecards and evolving standards from organizations including NIST, ISO, and IEEE, particularly mentioning the ISO 'Zero Trust AGI' [^][^]. Additionally, the 'Levels of AGI' framework offers an ontological structure to define AGI more precisely, clarifying ambiguous definitions by mapping performance and generality to specific benchmarks [^].
Resolving AGI claims in prediction markets typically relies on formal corporate announcements, not scientific validation. The resolution of prediction markets concerning when companies like OpenAI will achieve AGI usually hinges on public declarations from the companies themselves [^][^][^]. This dependency highlights a fundamental tension between rigorous scientific validation, corporate marketing objectives, and the absence of a universally accepted technical definition for AGI [^][^][^]. Current research does not explicitly detail how these specific scientific evaluation frameworks will be formally integrated into the resolution criteria for such prediction markets.
8. What technical hurdles and resource constraints support the more cautious AGI forecasts (post-2028) from experts like DeepMind's Demis Hassabis?
| Demis Hassabis AGI Forecast (Current) | within the next five years [^][^][^][^][^] |
|---|---|
| Resource Bottlenecks Severity | after 2028-2030 [^][^][^] |
| Key Technical Hurdles | continual learning, long-term reasoning, deeper memory, consistent agentic behavior [^][^][^][^][^] |
Cautious AGI forecasts often cite significant resource and technical challenges. Forecasts for Artificial General Intelligence (AGI) that extend into the 2030s are influenced by anticipated bottlenecks in resource availability and specific technical hurdles [^][^][^][^]. While these factors contribute to extended timelines, DeepMind's Demis Hassabis, who previously predicted AGI around 2030 in 2010, now foresees a high probability of its arrival within the next five years [^][^][^][^][^].
Resource limitations, particularly after 2028, pose significant AGI development obstacles. More conservative timelines (post-2028) for AGI development are notably impacted by resource constraints. These include a scarcity of high-quality training data, limitations in physical infrastructure such as energy and power grids, and insufficient chip manufacturing capacity, specifically from TSMC [^][^][^][^][^]. The logistical challenges of scaling compute and electricity supply are projected to become especially severe beyond 2028-2030 [^][^][^].
Critical technical hurdles extend beyond resources, affecting AGI's foundational capabilities. Beyond resource limitations, several critical technical impediments have been identified for AGI development. These encompass achieving robust continual learning, enhancing long-term reasoning and planning capabilities, developing deeper memory architectures that surpass current context windows, and ensuring consistent, reliable agentic behavior in real-world environments [^][^][^][^][^]. Additional challenges include diminishing returns from merely scaling existing transformer architectures and a widespread lack of consensus regarding appropriate AGI benchmarks [^][^][^][^][^][^][^].
9. What Could Change the Odds
Key Catalysts
Market predictions from Kalshi's "When will OpenAI achieve AGI?" offer outcomes such as "Before 2027", "Before 2028", and "Before 2030". This contract resolves if OpenAI announces attaining AGI by Dec 31, 2029 for the "Before 2030" option [^]. Similarly, Polymarket's "OpenAI announces it has achieved AGI before 2027?" market resolves to "Yes" only if OpenAI or an official representative announces AGI by Dec 31, 2026 (ET) [^].
Expert forecasts, reviewed by 80,000 Hours, report that as of Feb 2026, forecasters average about a 25% chance of AGI by 2029 and 50% by 2033, though they note definition and reliability issues [^] . 80,000 Hours also argues that the 2028–2032 window is particularly pivotal for determining whether progress accelerates toward AGI (~around 2030) or slows significantly, based on drivers like compute, investment, and algorithmic research [^]. OpenAI's own "Planning for AGI and beyond" document states that the AGI transition is expected to be gradual and could happen soon or far in the future, depending on takeoff speed, without committing to a specific date [^].
Key Dates & Catalysts
- Expiration: January 01, 2027
- Closes: January 01, 2030
10. Decision-Flipping Events
- Trigger: Market predictions from Kalshi's "When will OpenAI achieve AGI?" offer outcomes such as "Before 2027", "Before 2028", and "Before 2030".
- Trigger: This contract resolves if OpenAI announces attaining AGI by Dec 31, 2029 for the "Before 2030" option [^] .
- Trigger: Similarly, Polymarket's "OpenAI announces it has achieved AGI before 2027?" market resolves to "Yes" only if OpenAI or an official representative announces AGI by Dec 31, 2026 (ET) [^] .
- Trigger: Expert forecasts, reviewed by 80,000 Hours, report that as of Feb 2026, forecasters average about a 25% chance of AGI by 2029 and 50% by 2033, though they note definition and reliability issues [^] .
12. Historical Resolutions
No historical resolution data available for this series.
