Anthropic launched Claude Fable 5 in early June; because its capabilities were so powerful, it was banned by the US government shortly after launch, with national security clearance requirements mandating that all foreign access to Fable 5 and Mythos 5 be suspended—whether the person is inside or outside the United States, and including even Anthropic’s own foreign employees.
Many people are surely wondering whether China will soon have a model at the level of Fable 5. Elon Musk recently made a prediction on social media, saying it would “probably be Q1 next year.” Unexpectedly, the founder of Z.ai quickly replied to the post, saying, “It won’t take that long,” clearly hinting that the pace at which China’s frontier AI models are catching up may be faster than the outside world expects—perhaps we’ll see one before the end of this year.
Musk Predicts China’s Fable 5-Level AI Won’t Arrive Until Q1 Next Year; Z.ai Founder Responds: No Need to Wait That Long
Fable 5 is Anthropic’s latest publicly released flagship model; the company says it is a Mythos-class model tuned for general use and is more capable than any previously released Claude model, delivering very strong performance in areas such as software engineering, knowledge work, vision, and scientific research.
The main difference from Mythos 5 is that Fable 5 adds stricter security safeguards, with sensitive requests related to cybersecurity, biology, chemistry, or model distillation being handled by Claude Opus 4.8 instead.
Even so, the U.S. government quickly asked Anthropic to suspend access to Fable 5.
Recently, Teortaxes, an analyst on X who has long followed AI developments, mentioned in a post predicting the debut timeline of China’s Fable 5-level model that he believes the gap between GLM 5.2 and Fable-level models has narrowed to about seven months, roughly reaching the level of Claude Opus 4.7–4.8. This also means that China’s full version of Mythos (i.e., “Fable”) is expected to launch in November–December 2026.
However, Musk disagrees with this timeline; he believes it will arrive as early as Q1 2026.
Seeing the Q1 timeline, many people probably think it is already optimistic enough—after all, it has been moved up from Q4 to Q1—but for Chinese AI companies, this timing is still too late.
Shortly after, the CEO of Chinese company Z.ai replied below: “It won’t take that long.” This also means that a powerful Fable-level model could emerge in China as early as the end of this year:
Musk believes that while it may indeed be possible based on benchmark tests alone, if measured by genuine real-world utility, it would already be a very impressive achievement if China could pull it off by Q1 next year.
Either way, if Z.ai is indeed already developing a more powerful frontier AI model capable of rivaling—or even surpassing—Fable 5, it is something worth watching closely for the AI market. When that happens, it would not only signal another advance in the capabilities of Chinese frontier AI models, but could also intensify global competition among AI models, giving users more choices.
Source: KOCPC Chinese