Beijing-based Z.ai has released GLM 5.2, a large language model that claims near-parity with the most advanced US systems from Anthropic and OpenAI. The launch, which occurred just one day after Washington imposed and then lifted export controls on Anthropic’s Fable 5 and Mythos models, underscores the intensifying technological rivalry between the two powers.
According to Z.ai, GLM 5.2 performs almost on par with Anthropic’s Claude Opus 4.8 and OpenAI’s GPT 5.5. It operates on a 1 million token context window, equivalent to roughly 750,000 words of working memory, enabling it to handle extended coding tasks without losing coherence. The company says the model maintains “quality across long, messy coding-agent trajectories.”
Benchmark Performance
Z.ai tested GLM 5.2 on three benchmarks of complex, long-duration coding work. On open-ended technical projects lasting hours to days, it trails Claude Opus 4.8 by just 1% while edging past GPT 5.5 and Claude Opus 4.7. On a test measuring how well it can improve smaller models using a single GPU, it beats both GPT 5.5 and Opus 4.7, ranking second only to Opus 4.8. On the toughest marathon-length engineering tasks—such as building compilers—it still trails Opus 4.8 by 13%, though it remains second-best overall, according to the company. Across all three benchmarks, GLM 5.2 is the leading open-source model.
The open-source nature of GLM 5.2 is a key differentiator. Z.ai notes that the model has “no regional limits [and] technical access without borders,” meaning it can be modified for any purpose, including changing its output and sharing it freely. This contrasts sharply with the closed-source models from Anthropic and OpenAI, where users depend on the provider and cannot make alterations.
The timing of the release is notable. The US had briefly banned Anthropic from supplying its most advanced models to non-Americans, a move that was lifted on June 30. That episode highlighted the volatile export control landscape that European developers and businesses must navigate. For European companies and researchers, GLM 5.2’s open-source availability offers an alternative to US-dominated platforms, potentially reducing reliance on American cloud providers.
The broader AI race between the US and China has direct implications for Europe. While Washington has sought to limit China’s access to advanced semiconductors, Beijing has pursued a different path with cheaper, open-source models. This strategy could reshape the global AI supply chain, affecting everything from automotive software to healthcare diagnostics. European policymakers are watching closely, as the EU’s own AI Act and digital sovereignty goals may be influenced by the availability of open-source alternatives.
Z.ai’s claims, if independently verified, suggest that Chinese AI models are closing the gap with US leaders. The company’s focus on coding performance aligns with the needs of European tech hubs in Berlin, Paris, and Stockholm, where developers increasingly rely on AI-assisted tools. However, questions remain about data privacy and security when using models hosted outside the EU, especially given the bloc’s strict GDPR rules.
The release of GLM 5.2 also comes amid broader trade tensions. The EU is currently negotiating with China over electric vehicle tariffs and market access, with a deadline looming this autumn. As EPP Chief Manfred Weber warned, failure to reach a deal could escalate into a full-blown trade conflict. In this context, AI models become not just technological tools but strategic assets in a multipolar world.
For European readers, the key takeaway is that the AI landscape is no longer a duopoly between US giants. Chinese open-source models like GLM 5.2 offer a viable third path, but they also raise questions about data sovereignty and regulatory alignment. As the EU crafts its own AI strategy, the availability of such models may influence everything from research funding to procurement decisions in public administration.
