glm-5-turbo vs gpt-5.4-mini-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$0.75
Output$4.5
Latency (p50)4.9s
Output Limit128K
Function Calling
JSON Mode
-
InputText, Image
OutputText
in$0.75out$4.5cache$0.075–