Vercel Adds AI Model with Double the Throughput

Vercel's AI Gateway now includes the GLM 5.2 Fast model, offering developers a significant 2x performance boost for their AI applications.

Vercel has announced the availability of a new large language model, GLM 5.2 Fast, on its AI Gateway platform. The model is served through a specialized infrastructure called Wafer, which Vercel claims significantly boosts performance. According to the company's internal benchmarks, this new offering delivers twice the throughput compared to other serverless providers serving the same GLM model. Throughput, in this context, refers to the speed and efficiency of processing requests, meaning the model can handle more concurrent users or generate responses more quickly. Vercel reported that this performance advantage holds true across a range of common use cases, including tasks involving small and large amounts of text as well as those requiring the model to use external tools. This integration aims to provide developers with a high-performance option for building demanding AI-powered features directly within the Vercel ecosystem.

For developers, CTOs, and businesses, this update is significant because performance is a critical factor in the user experience of AI applications. A 2x improvement in throughput can mean the difference between a responsive, engaging chatbot and one that feels slow and frustrating. Faster model inference allows applications to serve users with lower latency, which is crucial for real-time interactions. This efficiency can also translate into lower operational costs, as teams can potentially handle a larger volume of requests with the same amount of resources. By offering a faster version of the GLM 5.2 model, Vercel makes its AI Gateway a more compelling choice for building production-ready applications that need to scale effectively. It provides engineering teams with another powerful tool to build sophisticated features without needing to manage complex AI model-serving infrastructure themselves.

This move highlights a growing trend in the cloud infrastructure market where competition is shifting from simply providing access to AI models to optimizing their delivery. As more models become available, the performance of the underlying serving layer is becoming a key differentiator for platforms like Vercel. Companies are increasingly looking for integrated solutions that simplify the entire development and deployment lifecycle, from front-end code to back-end AI inference. Vercel's partnership to serve GLM 5.2 Fast via Wafer strengthens its position as an all-in-one platform for modern web and AI development. For teams evaluating their technology stack, this means considering not just the capabilities of a specific AI model, but also the performance and efficiency of the platform that serves it.