From e1bc001872985a23af65c367b802ff8fb44edafc Mon Sep 17 00:00:00 2001
From: JoyFuture <35593546+JoyFuture@users.noreply.github.com>
Date: Fri, 8 May 2026 13:40:08 +0800
Subject: [PATCH] fix(mimo_v2): auto-disable multimodal when vision/audio
 configs are absent (#24652)

---
 python/sglang/srt/configs/model_config.py | 10 ++++++++++
 1 file changed, 10 insertions(+)

diff --git a/python/sglang/srt/configs/model_config.py b/python/sglang/srt/configs/model_config.py
index 4c2a7231e..0e62338f6 100644
--- a/python/sglang/srt/configs/model_config.py
+++ b/python/sglang/srt/configs/model_config.py
@@ -226,6 +226,16 @@ class ModelConfig:
                 logger.info(
                     f"Multimodal is disabled for {self.hf_config.model_type}. To enable it, set --enable-multimodal."
                 )
+            elif self.hf_config.architectures[0] in MIMO_V2_MULTIMODAL_ARCHS and not (
+                hasattr(self.hf_config, "vision_config")
+                and hasattr(self.hf_config, "audio_config")
+            ):
+                enable_multimodal = False
+                logger.info(
+                    "Multimodal is disabled for this MiMoV2 checkpoint: "
+                    "vision_config/audio_config not found in the model config "
+                    "(likely a text-only MiMoV2 variant)."
+                )
             else:
                 enable_multimodal = True