From 0f10fe6ce0a48d5a9825db9aa44f6dca2c5fa150 Mon Sep 17 00:00:00 2001
From: Nigel Jones <jonesn@uk.ibm.com>
Date: Fri, 6 Mar 2026 09:08:19 +0000
Subject: [PATCH] fix: ensure enough tokens for structured output in vLLM test
 (#591)

---
 test/backends/test_openai_vllm.py | 1 +
 1 file changed, 1 insertion(+)

diff --git a/test/backends/test_openai_vllm.py b/test/backends/test_openai_vllm.py
index f4e0e6466..34528861e 100644
--- a/test/backends/test_openai_vllm.py
+++ b/test/backends/test_openai_vllm.py
@@ -222,6 +222,7 @@ class Answer(pydantic.BaseModel):
         actions=[CBlock(value=prompt) for prompt in prompts],
         format=Answer,
         ctx=m_session.ctx,
+        model_options={ModelOption.MAX_NEW_TOKENS: 256},
     )
 
     assert len(results) == len(prompts)