Resolution criteria

This market will resolve to YES if, as of December 31, 2030, the leading frontier AI model developed primarily by a China-headquartered entity achieves parity with or surpasses the leading frontier AI model developed primarily by a US-headquartered entity.

Resolution will be determined sequentially using the following sources:

1. Epoch Capabilities Index (ECI): If Epoch AI is active and tracking capabilities as of December 31, 2030, this market resolves to YES if the highest-rated model developed by a China-headquartered organization has an ECI score equal to or higher than the highest-rated model developed by a US-headquartered organization.

2. LMSYS Chatbot Arena Leaderboard: If the ECI is unavailable, this market resolves to YES if the top model developed by a China-headquartered organization on the overall LMSYS Chatbot Arena Leaderboard has an Arena Elo rating equal to or higher than the top model developed by a US-headquartered organization as of December 31, 2030.

3. Stanford AI Index / CSET Reports: If both of the above metrics are defunct, the market will resolve to YES if the closest Stanford AI Index Report (e.g., the 2031 report tracking 2030 progress) or research from Georgetown's Center for Security and Emerging Technology (CSET) explicitly states that Chinese frontier models have matched or overtaken US frontier models in capability.

4. Creator Discretion: If none of the above sources are available or clear, the creator will resolve the market based on broad industry consensus of frontier AI capability rankings at the end of 2030.

Background

The geopolitical competition over artificial intelligence has increasingly centered on frontier large language and multimodal models. For several years, US labs (such as OpenAI, Anthropic, Google, and xAI) held a clear lead in overall capability. However, the performance gap has progressively narrowed.

According to the Stanford 2026 AI Index Report, the performance gap between the top US and Chinese models narrowed to just 2.7% as of March 2026 (comparing models like Anthropic's Claude 4.6 and ByteDance's Dola-Seed or DeepSeek's R1). While US hyperscalers outspend Chinese firms on raw private investment, Chinese labs have demonstrated highly efficient training methods and rapid capability leaps, leaving the long-term leader of the frontier AI race highly contested.