models
Added promptCacheTtl and subagentPromptCacheTtl settings for prompt cache control
10 de octubre de 2026 · Fuente original
New settings allow API-key and cloud-provider users to configure separate prompt cache TTLs: 1 hour for main conversation and 5 minutes for subagents, enabling better cache management and cost optimization.
New settings allow API-key and cloud-provider users to configure separate prompt cache TTLs: 1 hour for main conversation and 5 minutes for subagents, enabling better cache management and cost optimization.