From 9bbd646f40873cb4d62e30d041ea72875062b134 Mon Sep 17 00:00:00 2001
From: Novice <857526207@qq.com>
Date: Thu, 29 May 2025 09:49:44 +0800
Subject: [PATCH] fix: inner invoke llm token too long (#20391)

---
 api/core/plugin/backwards_invocation/model.py | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/api/core/plugin/backwards_invocation/model.py b/api/core/plugin/backwards_invocation/model.py
index 5ec9620f22..17cfaf2edf 100644
--- a/api/core/plugin/backwards_invocation/model.py
+++ b/api/core/plugin/backwards_invocation/model.py
@@ -58,6 +58,7 @@ class PluginModelBackwardsInvocation(BaseBackwardsInvocation):
                         LLMNode.deduct_llm_quota(
                             tenant_id=tenant.id, model_instance=model_instance, usage=chunk.delta.usage
                         )
+                    chunk.prompt_messages = []
                     yield chunk
 
             return handle()
@@ -68,7 +69,7 @@ class PluginModelBackwardsInvocation(BaseBackwardsInvocation):
             def handle_non_streaming(response: LLMResult) -> Generator[LLMResultChunk, None, None]:
                 yield LLMResultChunk(
                     model=response.model,
-                    prompt_messages=response.prompt_messages,
+                    prompt_messages=[],
                     system_fingerprint=response.system_fingerprint,
                     delta=LLMResultChunkDelta(
                         index=0,