From c2408889d2840ab9cd5eb87760a234bd62824038 Mon Sep 17 00:00:00 2001
From: UnaiAlias <52742669+UnaiAlias@users.noreply.github.com>
Date: Fri, 31 Jul 2026 18:53:57 +0000
Subject: [PATCH] fix: remove duplicate litellm stream call in
 fetch_response_litellm_openai

The streaming path called litellm.acompletion twice with the same kwargs.
The first result was assigned to a variable that is never read, and only
the second call's stream was returned. Every streamed request therefore
consumed two provider requests (e.g. two NVIDIA NIM requests) with the
same API key, halving effective per-key rate limits and leaking the
first stream. Keep a single acompletion call for the streamed path; same
fix in the tool_call_id truncation retry branch.
---
 src/cai/sdk/agents/models/chatcompletions/litellm_adapter.py | 2 --
 1 file changed, 2 deletions(-)

diff --git a/src/cai/sdk/agents/models/chatcompletions/litellm_adapter.py b/src/cai/sdk/agents/models/chatcompletions/litellm_adapter.py
index 795d65a0..8533bb68 100644
--- a/src/cai/sdk/agents/models/chatcompletions/litellm_adapter.py
+++ b/src/cai/sdk/agents/models/chatcompletions/litellm_adapter.py
@@ -68,7 +68,6 @@ async def fetch_response_litellm_openai(
     """
     try:
         if stream:
-            ret = await litellm.acompletion(**kwargs)
             stream_obj = await litellm.acompletion(**kwargs)
             return _build_response_obj(model_name, model_settings, tool_choice, parallel_tool_calls), stream_obj
         else:
@@ -102,7 +101,6 @@ async def fetch_response_litellm_openai(
             kwargs["messages"] = messages
 
             if stream:
-                ret = await litellm.acompletion(**kwargs)
                 stream_obj = await litellm.acompletion(**kwargs)
                 return _build_response_obj(model_name, model_settings, tool_choice, parallel_tool_calls), stream_obj
             else: