AGI in 5 Years? Manifold Bettors Put It at 62% After the GPT-6 Astra Launch. A Manifold Markets contract asking “Will AGI be achieved in the next 5 years?” is trading at a 62% implied probability of YES as of September 19, 2026, one of the busiest questions on the platform this week. The jump follows a compressed run of frontier model releases in early September and a widely quoted line from OpenAI president Greg Brockman: “Welcome to the AGI era.” Here is what the money is saying, what is actually driving it, and why the number is smaller than it looks.
What the Manifold market is actually asking
The contract, created by a user named kalb, is titled “Will AGI be achieved in the next 5 years?” on Manifold Markets, a play-money forecasting platform. The resolution rule matters more than the headline. The market resolves YES only if there is a “general consensus of those who interact with the potential AGI is that it is AGI” by its close date of January 19, 2029. If no such consensus forms, it resolves NO.
Two details are worth flagging up front. First, the “5 years” was measured from the market’s creation in early 2024, so the operative deadline is January 19, 2029, roughly 28 months out, not five years from today. Second, the resolution bar is social, not technical: it hinges on broad agreement that a system qualifies as AGI, using that specific term, rather than any single benchmark score. The creator notes that GPT-4 never reached that kind of consensus despite scattered claims at the time.
The odds as of September 19, 2026, and what they imply
As of September 19, 2026, the market sits at a 62% implied probability of YES on Manifold, with more than 100 traders holding positions and a fresh surge of activity over the prior 24 hours that pushed it into the platform’s top tier of active questions. On a play-money venue, that price reflects trader belief and attention rather than deep, real-money conviction, so it should be read as a sentiment gauge. For readers new to the mechanics, our explainer on how prediction markets work covers how these venues turn crowd bets into a probability estimate.
Taken literally, 62% says the crowd thinks a broadly recognized AGI is more likely than not before January 2029. That is aggressive relative to most other markets and most expert surveys, and the gap is largely a product of how loosely this particular contract defines its terms. It is a useful signal of where the conversation has moved, not a probability you should treat as calibrated. To translate a percentage into plain odds, see our note on implied probability.
What is driving the move
The proximate cause is a cluster of frontier model launches inside a single week in early September 2026. Three of the largest labs shipped upgrades within roughly 72 hours, and the density of releases is itself the story traders are reacting to.
OpenAI announced GPT-6 Astra and moved it to general availability in the first days of September. Anthropic released Claude Fable 5.1 and its restricted-access twin Claude Mythos 5.1 on September 1, 2026, with Fable becoming generally available across Claude, Claude Code and Cursor, per VentureBeat. Anthropic said the update cut typical workload costs by about 25% and highly agentic workload costs by as much as 45% through cheaper cached-context reads, alongside improved safeguards. Google DeepMind and Meta shipped frontier updates in the same window. When several labs release at once, forecasting crowds tend to read it as competitive acceleration, and that reading shows up in the price.
“Welcome to the AGI era”: the Astra launch
The single loudest catalyst was rhetorical. At the GPT-6 Astra briefing, OpenAI president Greg Brockman closed by telling reporters, “Welcome to the AGI era,” and suggested the milestone “might be about this model,” according to Axios. OpenAI described Astra as its most powerful model to date and emphasized its ability to operate browsers, spreadsheets and desktop software directly, framing it as a “generational leap” for cybersecurity, software engineering and science, as reported by Fortune.
The company had earlier defined AGI as a system that can perform all economically valuable work as well as or better than humans, a bar Astra does not clearly clear. As TechRepublic noted, the “AGI era” framing is a claim about direction of travel, not a declaration that AGI has arrived. Even so, executive language of this kind is exactly the sort of input short-horizon prediction markets react to.
The definition problem that makes these odds slippery
There is no single agreed definition of AGI, and that is the main reason AGI markets scatter so widely. Contracts tied to a company “announcing” AGI, or to a loose “consensus” test like this Manifold market, price higher than contracts demanding that a system match a skilled human across nearly all cognitive tasks reliably and at competitive cost. The stricter the definition, the lower the odds.
That divergence also separates traders from researchers. Prediction markets and lab leaders have clustered around late-2020s timelines, while broad expert surveys have historically pointed to much later dates. Benchmark branding adds to the confusion: strong results on tests labeled “ARC-AGI” measure specific reasoning skills, not general human-level competence, even though the name invites the leap. Read any AGI odds figure alongside the exact wording of what it resolves on.
How other markets price the same question
For a tech and finance reader, the useful comparison is across venues and timeframes. Real-money markets on Kalshi, a CFTC-regulated US exchange, track when OpenAI specifically will achieve AGI and price the near-term contracts well below Manifold’s headline.
| Market / question | Venue | Timeframe | Implied odds (approx.) |
|---|---|---|---|
| AGI achieved in next 5 years (consensus test) | Manifold (play money) | By Jan 19, 2029 | 62% YES (Sept 19, 2026) |
| OpenAI achieves AGI | Kalshi (regulated, real money) | Before 2028 | ~59% YES |
| OpenAI achieves AGI | Kalshi (regulated, real money) | Before 2030 | ~68% YES |
The pattern is consistent: the more the resolution rule leans on an “announcement” or “consensus” rather than a strict capability threshold, the higher the price. Kalshi’s near-term OpenAI contracts have risen through 2026 but still sit below the Manifold number, and both should be read next to their exact resolution language.
Who and what is involved
The named actors driving the narrative are the frontier labs and their executives. OpenAI and Greg Brockman supplied the headline quote and the flagship launch in GPT-6 Astra. Anthropic contributed Claude Fable 5.1 and Mythos 5.1, positioning the update around agentic coding and knowledge work with lower costs. Google DeepMind and Meta rounded out the release cluster. On the market side, Manifold hosts the play-money sentiment gauge, while Kalshi provides the regulated real-money reference point in the United States.
Tech-insider has tracked several of these threads in depth, including a head-to-head on GPT-6 Astra versus Gemini 3.8 and DeepSeek V4.1 Flash and a separate comparison of Muse Spark, GPT-6 Astra and Claude Fable. For how these forecasting venues behave in practice, see our coverage of a Manifold geopolitics market and a Kalshi odds surge.
What to watch next
Three things will move this contract between now and its January 2029 close. First, follow-on capability claims: whether Astra, Fable and their successors show sustained, independently verified autonomy on real economic work rather than benchmark wins. Second, executive language, since another “AGI era” style statement from a major lab tends to move short-horizon sentiment quickly. Third, safety pacing, including any evaluation pauses or incident disclosures, which cut both ways by signaling capability while raising deployment caution.
The cleaner read is to watch whether regulated real-money markets like Kalshi converge toward or diverge from Manifold’s play-money price. Sustained agreement across venues would be a stronger signal than a single platform spiking on launch-week attention.
Frequently asked questions
Does 62% mean AGI is likely by 2029? It means Manifold traders, on a play-money platform, currently price a better-than-even chance that a broadly recognized AGI emerges before January 19, 2029, under a loose consensus test. It is a sentiment reading, not a calibrated forecast, and it depends heavily on that definition.
Why is the Manifold number higher than Kalshi’s? The Manifold contract resolves on general “consensus” that a system is AGI, a looser bar than Kalshi’s company-specific AGI markets. Looser definitions and announcement-based triggers consistently price higher than strict capability thresholds.
Did OpenAI say GPT-6 Astra is AGI? No. Greg Brockman said “Welcome to the AGI era” and suggested the milestone “might be about this model,” which is a claim about trajectory, not a declaration that Astra meets OpenAI’s own definition of AGI.
Is Manifold real-money betting? No. Manifold uses play money, so its prices reflect attention and belief rather than capital at risk. Kalshi, by contrast, is a CFTC-regulated exchange using real money.
The Bottom Line
Sources
- Manifold Markets — Will AGI be achieved in the next 5 years?
- Axios — “Welcome to the AGI era,” OpenAI says as GPT-6 Astra debuts
- Fortune — OpenAI launches GPT-6 Astra, touts computer use
- TechRepublic — OpenAI Launches GPT-6 Astra as Brockman Says the AGI Era Has Arrived
- VentureBeat — Anthropic’s Claude Fable 5.1 and Mythos 5.1 arrive
- Kalshi — When will OpenAI achieve AGI?
Prediction markets carry risk and are not investment, financial, or betting advice. Prices reflect trader sentiment and can be wrong; play-money platforms such as Manifold do not involve capital at risk, while Kalshi is a CFTC-regulated US exchange. Availability is restricted by jurisdiction: Polymarket is not available to US persons, and access elsewhere varies. Participation is limited to those 18+ or 21+ as applicable in your location. If you or someone you know may have a gambling problem, call 1-800-GAMBLER for confidential help.
Related Coverage
- ARC-AGI-3 'Above Human' Bet Hits 92% on Manifold After GPT-6 Astra
- Will Polymarket IPO Before 2028? Manifold Bettors Now Put It at 43%
- Manifold Traders Give 'AGI Before 2027' Just a 2% Chance, Even as OpenAI Declares the 'AGI Era'
- 83% Odds: Manifold Bettors Say AI Will Play Any Game at Human Level by 2028
- Millennium Problems Bet: Manifold Prices 99% Odds Two Fall by 2040 After OpenAI's Navier-Stokes Claim