diff --git a/haystack/utils/openai_utils.py b/haystack/utils/openai_utils.py
index 40d3afbc7..fdde7306c 100644
--- a/haystack/utils/openai_utils.py
+++ b/haystack/utils/openai_utils.py
@@ -77,7 +77,7 @@ def _openai_text_completion_tokenization_details(model_name: str):
         if "text-davinci" in model_name:
             max_tokens_limit = 4097
             tokenizer_name = model_tokenizer
-        elif model_name.startswith("gpt-3.5-turbo-16k"):
+        elif model_name.startswith("gpt-3.5-turbo-16k") or model_name.startswith("gpt-35-turbo-16k"):
             max_tokens_limit = 16384
             tokenizer_name = model_tokenizer
         elif model_name.startswith("gpt-3"):
diff --git a/releasenotes/notes/support-azure-3.5-gpt-16k-model-ece0cfe03260748c.yaml b/releasenotes/notes/support-azure-3.5-gpt-16k-model-ece0cfe03260748c.yaml
new file mode 100644
index 000000000..642831b75
--- /dev/null
+++ b/releasenotes/notes/support-azure-3.5-gpt-16k-model-ece0cfe03260748c.yaml
@@ -0,0 +1,5 @@
+---
+fixes:
+  - |
+    gpt-35-turbo-16k model from Azure can integrate correctly
+