Dev Tools · 20h ago
Vercel Adds GLM 5.2 Fast Model to AI Gateway with 2x Throughput
Vercel's AI Gateway now supports GLM 5.2 Fast via Wafer, claiming 2x higher throughput than other providers. Benchmarks show 170+ tok/s for small contexts and 200+ tok/s for large contexts. The gateway offers unified API access, cost tracking, and zero data retention with no markup on pricing.
Meridian48 take
The performance claims are notable, but real-world gains depend on workload patterns and whether Wafer's architecture justifies switching from existing providers.
ai-gatewaymodel-deployment