glm-5-turbo vs gpt-5.4-nano-2026-03-17
Pricing, Performance & Features Comparison
Speed- and agent-optimized GLM-5 variant with low tool-call error rates for high-throughput agentic workloads.
Input$1.2
Output$4
Latency (p50)10.4s
Output Limit131K
Function Calling
JSON Mode
InputText
OutputText
in$1.2out$4cache$0.24–
Latency (24h)
Success Rate (24h)
Input$0.2
Output$1.3
Latency (p50)4.8s
Output Limit128K
Function Calling
JSON Mode
-
InputText, Image
OutputText
in$0.2out$1.3cache$0.02–