Fix phi4flash V1 compatibility

Author: aditchawdhary

tests/models/language/generation/test_hybrid.py

@@ -25,6 +25,7 @@
 
 HYBRID_MODELS = [
     "ai21labs/Jamba-tiny-dev",
+    "microsoft/Phi-4-mini-flash-reasoning",
     # skipping until vLLM implementation issues are resolved
     # "pfnet/plamo-2-1b",
     "Zyphra/Zamba2-1.2B-instruct",
@@ -37,6 +38,7 @@
 V1_SUPPORTED_MODELS = [
     "state-spaces/mamba-130m-hf",
     "ai21labs/Jamba-tiny-dev",
+    "microsoft/Phi-4-mini-flash-reasoning",
     "yujiepan/mamba2-codestral-v0.1-tiny-random",
     "Zyphra/Zamba2-1.2B-instruct",
     "hmellor/tiny-random-BambaForCausalLM",

tests/models/registry.py

@@ -282,7 +282,6 @@ def check_available_online(
     "Phi3ForCausalLM": _HfExamplesInfo("microsoft/Phi-3-mini-4k-instruct"),
     "Phi4FlashForCausalLM": _HfExamplesInfo("microsoft/Phi-4-mini-flash-reasoning", # noqa: E501
                                         trust_remote_code=True,
-                                        v0_only=True,
                                         max_model_len=10240),
     "PhiMoEForCausalLM": _HfExamplesInfo("microsoft/Phi-3.5-MoE-instruct",
                                          trust_remote_code=True),