Prompt caching

config.template.toml

@@ -110,6 +110,15 @@ api_key = "your-api-key"
 # Cost per output token
 #output_cost_per_token = 0.0
 
+# Discount to apply for cache hits (e.g., 0.8 for 80% discount)
+# Required if input_cost_per_token and output_cost_per_token are set
+#cache_hit_discount = 0.8
+
+# Premium to apply for cache writes (e.g., 0.2 for 20% premium)
+# Required if input_cost_per_token and output_cost_per_token are set
+# Only used for Anthropic models
+#cache_write_premium = 0.2
+
 # Custom LLM provider
 #custom_llm_provider = ""