The agent engineering platform.
Python agentsaiai-agentsanthropicchatgptdeepagentsenterpriseframeworkgeminigenerative-ailangchainlanggraphllmmultiagentopen-sourceopenaipydanticpythonragtypescript
anthropic: store cache ttl details on usage metadata (#31393)
ccurme0e3f35effe21
diff --git a/libs/partners/anthropic/langchain_anthropic/chat_models.py b/libs/partners/anthropic/langchain_anthropic/chat_models.pyindex 3bc75804f0..60a95747de 100644--- a/libs/partners/anthropic/langchain_anthropic/chat_models.py+++ b/libs/partners/anthropic/langchain_anthropic/chat_models.py@@ -2129,6 +2129,13 @@ def _create_usage_metadata(anthropic_usage: BaseModel) -> UsageMetadata:"cache_read": getattr(anthropic_usage, "cache_read_input_tokens", None),"cache_creation": getattr(anthropic_usage, "cache_creation_input_tokens", None),}+ # Add (beta) cache TTL information if available+ cache_creation = getattr(anthropic_usage, "cache_creation", None)+ if cache_creation:+ for k in ["ephemeral_1h_input_tokens", "ephemeral_5m_input_tokens"]:+ v = getattr(cache_creation, k, None)+ if v:+ input_token_details[k] = v# Anthropic input_tokens exclude cached token counts.input_tokens = (