China Cuts Its AI Gap to 6%: Moonshot and Zhipu Are Not a One-Off

On July 23, 2026, Bloomberg Intelligence reported that Chinese AI models narrowed their performance gap with American models to a record-low 6% in June 2026, down from 9% a month earlier.

How close has China come to the US?

The performance gap between Chinese and American AI models stood at 6% in June 2026, the lowest figure in Bloomberg Intelligence's observations. In May 2026 the gap was estimated at 9%, so it shrank by roughly a third in a single month. It is the narrowest gap the analysts have recorded in their series of measurements.

The figure is an aggregate estimate: the lower the percentage, the closer Chinese systems have come to the best American models on overall metrics. Bloomberg Intelligence is the research arm of Bloomberg, and the existence of separate estimates for May and June 2026 points to regular month-by-month tracking rather than a one-off snapshot.

• Performance gap: 6% in June 2026 (record low)
• One month earlier, in May 2026: 9%
• Reduction over the month: roughly one third
• Source of the estimate: Bloomberg Intelligence
• Strong Chinese players in focus: Moonshot AI and Zhipu

Why is this not a one-off?

The progress comes not from a single Chinese lab but from several at once, and that is the report's key argument. Bloomberg Intelligence singles out Moonshot AI and Zhipu by name: when more than one company posts strong results, the case for a random one-time leap looks weaker.

That is why the analysts read the June numbers not as an outlier but as a sign of a sustained trend. By the report's logic, China's AI development ecosystem is catching up with the American one on several fronts at once, and the drop from 9% to 6% in a month fits that picture rather than contradicting it.

What does this call into question?

Narrowing the gap to 6% raises doubts about how durable US technological superiority in AI really is, a direct conclusion of the Bloomberg Intelligence report. Until recently the lead of American labs was considered comfortable, but the measurements show it shrinking fast.

"The narrowing of the gap raises questions about the sustainability of US technological superiority in the sector," according to the

Bloomberg Intelligence report.

For the market, this is a signal: AI leadership is ceasing to be the monopoly of a few American companies and is becoming a matter of genuine international competition. Whether US leaders can hold their pace will affect both investment flows and the balance of power in the global model market.

What does this mean?

If the trend from 9% to 6% in a month continues, the gap between Chinese models and American ones could become statistically insignificant in the foreseeable future. For users and businesses, that means more competitive alternatives; for American labs, it means having to keep up the pace rather than rely on their former lead.

Frequently asked questions

How far behind are

Chinese AI models compared with American ones?

According to Bloomberg Intelligence, the performance gap was 6% in June 2026, a record low, compared with 9% in May 2026.

Why is the narrowing of the gap considered a sustained trend?

Because strong results are coming not from one Chinese company but from several, including Moonshot AI and Zhipu, which argues against the one-time-leap explanation.