OpenAI vs. Anthropic: First to announce recursive self-improvement?

Foretrace AI research

Understand the odds. Examine the reasoning.

Polymarket shows market prices. Foretrace adds AI probability estimates, opposing evidence and a price-dependent outcome or pass decision for each open market.

Choose a market to analyze 1 open question · Evidence required

24h volume$141.9

Total volume$1.2K

Liquidity$168.8

Event end date · UTCJan 1, 2028

Price history

Market probability

Loading price history…

Polymarket · UTC

Sampled market observations, not AI forecasts. Lines connect available observations. Legend values are current provider quotes; chart points may be older.

Outcomes & resolution rules

1 market

These are market-implied probabilities, not AI forecasts. A displayed quote may differ from the price available for your order.

OpenAI vs. Anthropic: First to announce recursive self-improvement?Accepting ordersMarket odds42.5%

OpenAI vs. Anthropic: First to announce recursive self-improvement?

OpenAI42.5%Polymarket estimate

Anthropic57.5%Polymarket estimate

Liquidity: $168.8
Deadline · UTC: Jan 1, 2028
OpenAI bid / ask: 36.0¢ / 49.0¢
Trading fees: Enabled · verify on Polymarket

Get the Foretrace AI forecast

See a probability estimate with uncertainty, the evidence for and against it, and a conditional decision at the available price.

Fresh successful report: 1 credit. Valid shared report: free.

Resolution rules

This market will resolve to "OpenAI" if OpenAI, or an official representative of OpenAI, announces that OpenAI has achieved recursive self-improvement before Anthropic does so. This market will resolve to "Anthropic" if Anthropic, or an official representative of Anthropic, announces that Anthropic has achieved recursive self-improvement before OpenAI does so. If neither company makes a qualifying announcement by December 31, 2027, 11:59 PM ET, or if both make qualifying announcements on the same calendar day, this market will resolve 50-50. A qualifying announcement by any other company does not end the race or affect resolution. Recursive self-improvement means a capability, as described in the company's announcement, in which the company's AI autonomously conducts the research and development needed to materially improve the general capabilities of its most capable or near-frontier AI systems, or to produce a more capable successor, with the improved system able to drive further such improvements. Humans may set high-level goals and provide resources, oversight, and approval, but must not supply the substantive research choices or the technical solution. Improvements that are toy, narrow, sandboxed, proof-of-concept, or limited to prompts, skills, or benchmark scores will not qualify. A qualifying announcement must describe the capability as currently achieved in an official capacity. Statements that AI is assisting, accelerating, or writing code for the company's research, statements of future plans or expectations, speculation, interviewer characterizations, leaks, model outputs, and third-party claims will not qualify. A public release, a named model, a technical demonstration, and independent verification are not required. The primary resolution source for this market will be official information from OpenAI and Anthropic and/or their official representatives; however, a consensus of credible reporting may also be used.