China's Zhipu closes gap on top US models as GLM-5.2 gains traction

Read the original article →

What happened

CNBC reported that Zhipu AI, the Beijing-based lab known as Z.ai, is narrowing the gap with leading US models through GLM-5.2, an open-weight model released under the permissive MIT license. Users highlighted its coding performance and low cost.

GLM-5.2 is a 744-billion-parameter mixture-of-experts model with a context window of up to 1 million tokens. It reportedly sits within about 1% of Anthropic's Opus 4.8 on a closely watched agentic benchmark while costing roughly one-fifth as much.

Why it matters

Observers described the release as a new 'DeepSeek moment' for cost-effective coding models. With Anthropic and OpenAI constrained by US government restrictions, an open, cheap, high-performing Chinese model has room to gain adoption.

The pattern suggests those same restrictions may be helping Chinese labs close the gap rather than widening the US lead.

MintedBrain take

An MIT-licensed model this close to the frontier is worth testing on your own coding workloads before assuming the US labs are the only serious option. Weigh the cost savings against your data-governance and provenance requirements.

References

This article was originally published at The Decoder. For the full piece, read the original article.

Discussion

  • Loading…

← Back to News