Z.ai Raises GLM 5.3 Prices on OpenRouter as NVIDIA Publishes Dual-DGX Serving Guide

Z.ai: GLM 5.3 now costs $0.084 per million input tokens and $3.6 per million output tokens on OpenRouter, up from $0.0699 per million input tokens and $1.4 per million output tokens on 2026-11-13. The input price rose 20.2 percent and the output price rose 157.1 percent. The model has 753.3 billion parameters and a 1048576 token context window.

Among other Z.ai models listed today, Z.ai: GLM 5.2 costs $0.007 per million input tokens and $6.8 per million output tokens, while Z.ai: GLM 4.7 Flash costs $0.0605 per million input tokens and $0.4 per million output tokens. Z.ai: GLM 5.3 Flash costs $0.0399 per million input tokens and $0.0784 per million output tokens. Z.ai: GLM 4.5 Air costs $0.2 per million input tokens and $1.1 per million output tokens, and Z.ai: GLM 4.6V costs $0.3 per million input tokens and $0.9 per million output tokens.

NVIDIA announced "When One GPU Isn’t Enough: Serving GLM-5.3 Across Two NVIDIA DGX Stations with vLLM". According to NVIDIA, GLM-5.3 NVFP4 is a quantized mixture-of-experts model with 753 billion parameters and 40 billion activated per token, requiring about 464 GB of GPU memory. According to NVIDIA, AIPerf benchmarks on the SPEED-Bench Coding dataset show overall output throughput peaking at 607.8 tokens per second at concurrency 16, falling to 385.0 tokens per second at concurrency 32 with mean request latency of 453.5 seconds. According to NVIDIA, two NVIDIA DGX Stations connected over ConnectX-8 RDMA serve the model through one OpenAI-compatible endpoint using tensor parallelism (TP=2), with each GPU holding and computing roughly half the weights.