Z AI's latest open model, GLM-5.2, has posted the highest score ever recorded by an open model on the Artificial Analysis Intelligence Index — a 51. The result, announced this week, places the model ahead of many proprietary rivals and signals a shift in the competitive landscape of large language models.
How the Index Works
The Artificial Analysis Intelligence Index ranks AI models on a range of tasks. A score of 51 puts GLM-5.2 at the top of the open model category. The benchmark is widely used to compare performance across both open and proprietary systems. Z AI did not disclose which specific tests contributed most to the score.
What GLM-5.2 Brings
Z AI says the model's advancements could redefine AI's role in complex coding and long-term tasks. Those are areas where proprietary models have traditionally held an edge. By releasing GLM-5.2 as an open model, the company is betting that the performance gap between open and closed systems is narrowing — and fast.
Market Dynamics at Play
The score challenges the assumption that only proprietary models can handle demanding, multi-step problems. That's a direct challenge to companies that charge for access to their systems. The release of GLM-5.2 could influence market dynamics by pushing down prices or forcing proprietary developers to improve performance to stay ahead. For now, the open model community has a new benchmark to beat.
The next milestone? Z AI has not announced a follow-up model, but with a score of 51 on the table, the pressure is now on both open and proprietary labs to respond.
