Within less than 12 months, four or five Chinese companies caught up to the frontier of the AI industry.

Judge as of 2026-01-07. The claim has three parts: four or five distinct Chinese companies, AI models at or near the industry frontier, and catching up within less than 12 months (roughly 2025). True: independent benchmark leaderboards or evaluations (e.g. LMArena, Artificial Analysis, Epoch AI, the Stanford AI Index) show that at least four Chinese companies (e.g. DeepSeek, Alibaba/Qwen, Moonshot/Kimi, Zhipu, MiniMax, ByteDance) released models during the previous ~12 months that ranked at or very close to the top US frontier models, for example within a few points on major aggregate benchmarks or in the top tier of leaderboards. Mostly true: the same picture holds but the count is a little off (three, or six or more), or the timeframe was a bit longer than 12 months. Mixed: Chinese models came close on some benchmarks or in open-weight categories but clearly trailed the overall frontier, or only two or three companies came close. Mostly false: only one company came near the frontier, or the evidence shows a clear, persistent gap that commentators described as months behind and the claim overstates that. False: no Chinese company's models were near the frontier at the time.

Set before the research began.

1. E1 Chinese AI models have lagged the US frontier by 7 months on average since 2023 Epoch AI · 2 January 2026

   “Since 2023, every model at the frontier of AI capabilities, as measured by the Epoch Capabilities Index, has been developed in the United States. Over that same period, Chinese models have trailed US capabilities by an average of seven months, with a minimum gap of four months and a maximum gap of 14.”

2. E2 Chinese AI models have lagged the US frontier by 7 months on average since 2023 Epoch AI · 2 January 2026

   “In practice, the first point where a Chinese model surpasses GPT-4 is May 2024 (a gap of 14 months), and no Chinese model has yet surpassed the ECI of OpenAI’s o3 model, released in April 2025.”

3. E3 Technical Performance Stanford HAI (AI Index 2026)

   “U.S. and Chinese models have traded places at the top of performance rankings multiple times since early 2025. In February 2025, DeepSeek-R1 briefly matched the top U.S. model. As of March 2026, the top U.S. model leads by 2.7%, with a gap that fluctuated over the past year while remaining in the single digits.”

4. E4 Competing AI strategies for the US and China | Brookings Brookings Institution · 10 June 2026

   “China’s top AI models continue to lag behind American frontier models by several months or more. American AI models maintain a clear lead in overall performance across a wide range of industry benchmarks, from math and reasoning to code generation and long-horizon agentic tasks.”

Epoch AI, writing days before the judge-as-of date, reports that every frontier model since 2023 was developed in the US, that Chinese models trailed by an average of seven months, and that no Chinese model had yet passed OpenAI's April 2025 o3 on its capabilities index (E1, E2). The Stanford AI Index 2026 reports that US and Chinese models traded places at the top several times from early 2025, with DeepSeek-R1 briefly matching the top US model and the gap staying in single digits (E3). A later Brookings piece also describes a lag of several months (E4). Together this shows Chinese models came close to the frontier and at times matched or traded places with top US models [E3], though on Epoch's capabilities index they trailed the US frontier by about seven months on average [E1, E2]. None of the evidence establishes that four or five distinct Chinese companies reached the frontier.