Elon Musk guessed first quarter of next year. That was his prediction for when a Chinese AI model would match Anthropic’s new Fable 5. Tang Jie, CEO of Beijing-based Zhipu, answered in public on social media: “won’t take that long.” The exchange, rare in its directness between a Chinese tech chief and an American billionaire, does more than settle a timeline bet.
It exposes the core tension in the US-China artificial intelligence race right now. Two different theories of what “better” means are colliding.
Musk’s argument is rooted in revenue. Anthropic, he pointed out, focuses on “maximizing useful intelligence” — and that shows up in how much money the company makes. Benchmarks, in his view, are not the same as real-world value.
A model can score high on tests and still fail to be genuinely useful to customers who pay. Tang Jie shot back with a line that sounds almost philosophical: “Focus is all we need, especially focusing on what intelligence truly is.” The Zhipu CEO did not mention revenue.
He did not mention market share. He went straight to the nature of intelligence itself. This is not a trivial disagreement.
It reflects how each side is betting its future. Zhipu’s latest model, GLM-5.2, already performs roughly on par with Anthropic’s Opus 4.7 and 4.8 — models launched in April and May this year. The company’s benchmarks show GLM-5.2 beating OpenAI’s GPT-5.5 consistently.
Those are hard numbers. They are not opinions.
But Musk is not wrong either. Anthropic’s Fable 5 is a different class of model. And “true usefulness” — his phrase — is harder to measure than a benchmark score.
A model that nails a test but fails in a live business workflow is not a product. It is a science project.
Tang Jie’s retort suggests he believes the gap between test performance and practical value is narrower than Musk thinks. Or that it will close faster. Or that the definition of “intelligence” itself is being contested, and Zhipu is working from a different premise.
State and international reports have noted a sharp rise in Chinese AI model capabilities recently. Officials are said to be watching developments closely. The exchange between the two CEOs has revived debate about the AI talent race between the US and China — a race that has been running for years but is now entering a more concrete phase.
What the exchange did not settle: whether China’s rapid benchmark gains will translate into products that generate real revenue and real adoption. Musk is betting they will not — at least not by Q1.
Tang Jie is betting they will, and sooner. One thing is clear. The days of Chinese AI labs being years behind are over.
GLM-5.2 matches a model from April. The next generation aims at Fable 5.
Tang Jie says that target will not take long to hit. He is the one building the model. He should know.
Whether that speed also delivers what Musk calls “true usefulness” is the question neither CEO answered. Both walked away from the exchange claiming their approach would win. The market — and the benchmarks — will decide which one was right.
{“@context”: “https://schema.org”, “@type”: “NewsArticle”, “headline”: “Chinese AI CEO Disputes Musk’s Timeline for Matching US Model”, “description”: “Elon Musk guessed first quarter of next year. That was his prediction for when a Chinese AI model would match Anthropic’s new Fable 5. Tang…”, “datePublished”: “2026-06-23T09:27:00+00:00T09:30:00+00:00”, “author”: {“@type”: “Person”, “name”: “Arjun Mehta”}, “publisher”: {“@type”: “Organization”, “name”: “InfoPulse Today”, “url”: “https://newsfeed.infopulsetoday.com/”}}
