Chinese startup Z.ai (also known as Zhipu AI) has launched GLM-5.2, a large language model built for complex coding projects that it says outperforms OpenAI's GPT-5.5 on key benchmarks.
According to The Economic Times, GLM-5.2 tops GPT-5.5 on those benchmarks while featuring a 1 million token context window — a measure of how much text and code the model can consider at once. A larger context window lets a model hold more of a project in view, which matters for software work that spans many files.
BizzBuzz reports that GLM-5.2 beats GPT-5.5 on coding benchmarks at one-sixth the cost. If that pricing holds up in real use, it would make the model a notably cheaper option for developers running heavy coding workloads, where costs add up quickly.
The release also reflects a broader strategic bet. According to Futu (富途牛牛), Zhipu AI is wagering on long-context capabilities and on domestic computing-power ecosystems as it tries to join the global elite of large-model makers — a nod to China's push to build competitive AI while relying less on foreign hardware.
It's worth noting that the performance claims come from benchmark results highlighted around the launch rather than independent, long-term testing, so how the model performs across varied real-world projects remains to be seen.
Why it matters: a Chinese model claiming to match a top U.S. system on coding at a fraction of the price signals intensifying global competition in AI — and could pressure pricing and adoption far beyond China.