diff --git a/vllm_omni/entrypoints/openai/serving_speech.py b/vllm_omni/entrypoints/openai/serving_speech.py
index f22f2a3d8..15579e4e0 100644
--- a/vllm_omni/entrypoints/openai/serving_speech.py
+++ b/vllm_omni/entrypoints/openai/serving_speech.py
@@ -77,6 +77,7 @@ _TTS_MODEL_STAGES: set[str] = (
     | _MING_TTS_MODEL_STAGES
     | _MOSS_TTS_MODEL_STAGES
 )
+_SAMPLING_MAX_TOKENS_TTS_MODEL_TYPES = {"fish_tts", "qwen3_tts", "voxtral_tts", "cosyvoice3", "voxcpm2"}
 _TTS_LANGUAGES: set[str] = {
     "Auto",
     "Chinese",
@@ -876,6 +877,11 @@ class OmniOpenAIServingSpeech(OpenAIServing, AudioMixin):
         if self._tts_model_type == "voxcpm":
             return self._validate_voxcpm_request(request)
         if self._tts_model_type == "voxcpm2":
+            if request.max_new_tokens is not None:
+                if request.max_new_tokens < _TTS_MAX_NEW_TOKENS_MIN:
+                    return f"max_new_tokens must be at least {_TTS_MAX_NEW_TOKENS_MIN}"
+                if request.max_new_tokens > _TTS_MAX_NEW_TOKENS_MAX:
+                    return f"max_new_tokens cannot exceed {_TTS_MAX_NEW_TOKENS_MAX}"
             return None  # VoxCPM2 accepts any text input
         if self._tts_model_type == "ming_flash_omni_tts":
             return self._validate_ming_tts_request(request)
@@ -1859,13 +1865,23 @@ class OmniOpenAIServingSpeech(OpenAIServing, AudioMixin):
             sampling_params_list[0].extra_args.update(request.extra_params)
             logger.info("Applied extra_params: %s", request.extra_params)
 
-        # Fish defaults come from stage_configs YAML. Only override when the caller
-        # explicitly requests a different generation length.
-        if self._is_fish_speech and request.max_new_tokens is not None and sampling_params_list:
+        # Some TTS model defaults come from deploy YAML. Their AR
+        # generation length is controlled by SamplingParams.max_tokens, so only
+        # override it when the caller explicitly requests max_new_tokens.
+        if (
+            self._tts_model_type in _SAMPLING_MAX_TOKENS_TTS_MODEL_TYPES
+            and request.max_new_tokens is not None
+            and sampling_params_list
+        ):
             import copy
 
             sampling_params_list = copy.deepcopy(sampling_params_list)
             sampling_params_list[0].max_tokens = request.max_new_tokens
+            if self._tts_model_type == "cosyvoice3":
+                sampling_params_list[0].min_tokens = min(
+                    getattr(sampling_params_list[0], "min_tokens", 0),
+                    request.max_new_tokens,
+                )
 
         # Propagate per-request seed to sampling params so both Slow AR
         # and Fast AR produce deterministic output for the same seed.