Will a Chinese company have a top ___ AI model by December 31?
Chinese labs are competitive enough that a top-3 breakthrough is plausible, but the leaderboard has recently been dominated by Anthropic, OpenAI, and Google. I think the market is slightly too pessimistic, though the path to Yes still requires one Chinese model to beat several entrenched frontier systems.
Analysis
The latest ranking snapshots point to a clear near-term pattern: the very top of the overall leaderboard is still controlled by a small set of U.S.-based frontier labs, with Anthropic, OpenAI, and Google most often occupying the first three positions. That makes a Chinese top-3 finish look unlikely at the moment, and the current market price of about 22.5% reflects that skepticism reasonably well. Still, the market is not asking whether a Chinese model is the best model in the world for the whole year; it only needs one Chinese company to break into the top three at any check during the year, which is a meaningfully easier target.
The strongest case for Yes is that Chinese companies continue to iterate quickly and have already shown they can produce highly competitive frontier models. DeepSeek, Z.AI, Alibaba’s Qwen line, ByteDance, Tencent, MiniMax, and others are all active, and the recent cadence of releases suggests that a single strong launch or a favorable leaderboard refresh could move one of them into the top tier. In a market like this, short-lived leaderboard leadership matters just as much as durability, so a temporary gap closing by a Chinese model is enough to settle the question positively.
The strongest case for No is that the current leaderboard structure seems to reward the very best proprietary frontier systems, and the incumbents have both the research depth and the deployment scale to defend their positions. Even if Chinese models keep improving, moving from “competitive and often near the top” to “strict top 3 overall” requires overtaking multiple models from labs that are also advancing rapidly. Since the market resolves based on a strict score ranking with ties excluded, Chinese models need a clean, unambiguous climb into the top three, which is harder than simply being in the conversation.
Overall, I lean slightly above the current market-implied probability because Chinese labs have enough momentum and release velocity to make a top-3 appearance realistic before year-end. But the base rate still favors the established leaders, so this remains a minority outcome rather than a coin flip.
Arguments
For
- Arguments for Yes: Chinese labs are iterating fast enough that one strong release could push a model into the top 3 before December 31.
- Arguments for Yes: The market only requires a temporary top-3 appearance, which is easier than ending the year in that position.
Against
- Arguments against Yes: The current top of the leaderboard is still dominated by Anthropic, OpenAI, and Google, leaving little room for others.
- Arguments against Yes: Chinese models are competitive but still appear more often as leading open-weight contenders than as strict top-3 overall models.
Key drivers
- Rapid Chinese model releases increase the chance of a short-lived top-3 breakthrough.
- The current leaderboard is still anchored by Anthropic, OpenAI, and Google at the very top.
- A single strong update can move a model into the top 3 because the market checks the leaderboard over time, not just at year-end.
Risk factors
- Frontier U.S. labs may continue improving quickly enough to keep Chinese models outside the top 3.
- Leaderboard methodology changes or score volatility could prevent a Chinese model from holding a clean top-3 ranking.
- Chinese models may remain highly competitive without crossing the narrow threshold needed for a strict top-3 placement.
Scenarios
Best case
A major new Chinese release or a favorable leaderboard update lifts a Chinese model into the top 3, and it stays there long enough to be clearly observed at a check time.
Most likely
Chinese models stay near the frontier and occasionally close the gap, but Anthropic, OpenAI, and Google remain the most likely occupants of the top three positions.
Worst case
The leading U.S. frontier labs keep incrementally improving, and Chinese models remain just outside the top 3 despite continued progress.
More from this day
- CompaniesKalshi1y
Starbucks total global stores in 2026
AI96%MKT9%Edge+87Hidden GemStarbucks is already above the 41,800-store threshold in its latest reported results, and management’s 2026 opening guidance makes an above-threshold 2026 report highly likely. I put the Yes probability at 96%.
- PoliticsKalshi1y
2026: Trump's bad year?
AI68%MKT10%Edge+58Hidden GemTrump already has multiple concrete 2026 setbacks in court, so the bear-case narrative looks more likely than not. The market’s 11% price appears too low unless the event definition is far narrower than the news flow suggests.
- PoliticsKalshi2y
Who will Trump pardon?
AI3%MKT47%Edge-44HypedBarron Trump receiving a presidential pardon by Jan. 21, 2029 looks extremely unlikely absent some new federal legal exposure. My independent estimate is 3%, far below the current market price.