From 88a9cdd4397932aa189b9999b346f5cde427ef1e Mon Sep 17 00:00:00 2001 From: Wentao Ye <44945378+yewentao256@users.noreply.github.com> Date: Tue, 16 Jun 2026 12:31:32 -0400 Subject: [PATCH] [Model Runner V2] Enable GraniteMOE for MRv2 by default (#45461) Signed-off-by: yewentao256 --- tests/test_config.py | 2 +- vllm/config/vllm.py | 1 + 2 files changed, 2 insertions(+), 1 deletion(-) diff --git a/tests/test_config.py b/tests/test_config.py index 918f89beb8..e6c122de83 100644 --- a/tests/test_config.py +++ b/tests/test_config.py @@ -168,7 +168,7 @@ def test_v2_model_runner_env_tri_state(monkeypatch, env_value, expected): is_moe=True, is_quantized=False, ), - False, + True, ), ( SimpleNamespace( diff --git a/vllm/config/vllm.py b/vllm/config/vllm.py index 98ec40e860..3eae4c2cd1 100644 --- a/vllm/config/vllm.py +++ b/vllm/config/vllm.py @@ -70,6 +70,7 @@ DEFAULT_V2_MODEL_RUNNER_ARCHITECTURES = frozenset( "Qwen3ForCausalLM", "DeepseekV2ForCausalLM", "Qwen2MoeForCausalLM", + "GraniteMoeForCausalLM", "LlamaForCausalLM", "MistralForCausalLM", }