chore(examples): bump module_client.py to gpt-5.5

The module-level client example still uses bare gpt-4. gpt-5.5 is the
current default text and reasoning model per the openai-cookbook
("use gpt-5.5 for the strongest code review accuracy") and the
openai-agents-python docs (voice/quickstart.md, reasoning_content
example), and is documented as the latest default in
openai/codex's latest-model.md.

The model: parameter is typed Union[str, ChatModel] so this works
even though Stainless has not yet synced gpt-5.5 into the ChatModel
literal here. The example demonstrates module-level client
configuration, the model choice is incidental.

Author: lucky-verma

examples/module_client.py

@@ -9,7 +9,7 @@
 
 # all API calls work in the exact same fashion as well
 stream = openai.chat.completions.create(
-    model="gpt-4",
+    model="gpt-5.5",
     messages=[
         {
             "role": "user",