fix: ZONOS2 only generating 1024 tokens max

1 files changed, 11 insertions, 0 deletions

diff --git a/app/tests/test_tts_sglomni.py b/app/tests/test_tts_sglomni.py
index 2dee364..58ca57c 100644
--- a/app/tests/test_tts_sglomni.py
+++ b/app/tests/test_tts_sglomni.py
@@ -248,6 +248,17 @@ class PayloadTests(unittest.TestCase):
         client._seed = 7
         self.assertEqual(client._request_payload("Hello.")["seed"], 7)
 
+    def test_zonos2_payload_raises_the_generation_cap(self):
+        """Zonos2's 1024-frame engine default caps a request at ~12 s."""
+        client = self._make_client("zonos2")
+        payload = client._request_payload("Hello.")
+        self.assertEqual(payload["max_new_tokens"], 12288)
+
+    def test_models_without_a_cap_send_no_max_new_tokens(self):
+        client = self._make_client("higgs_audio_v3_tts")
+        self.assertNotIn("max_new_tokens",
+                         client._request_payload("Hello."))
+
 
 class RequestErrorTests(unittest.TestCase):
     """OpenAI-style error envelopes decide retryability."""