Zhipu's GLM 5.2 artificial intelligence model has generated significant attention, comparable to the buzz surrounding DeepSeek's launch last year. This new open-source model has achieved performance levels close to Anthropic's Opus 4.8 on a key benchmark, while being available at approximately one-fifth of the cost.
The rapid increase in developer engagement, evidenced by the surge in OpenRouter token traffic, indicates strong market interest. Unlike DeepSeek, which was ultimately viewed as a temporary novelty, GLM 5.2 excels in practical applications such as planning, coding, and testing—areas where businesses are eager to enhance automation.
As companies face rising AI costs, the focus on 'intelligence per dollar' makes Zhipu's model particularly appealing. Gabe Pereyra, co-founder of Harvey, noted the surprising pace at which open-source models have advanced, with GLM 5.2 being the first to compete effectively with established closed-source models.
Additionally, the current climate of federal oversight affecting companies like Anthropic and OpenAI, which have faced restrictions on their models, positions Zhipu's freely accessible model as a more stable option for enterprises seeking AI solutions