diff --git a/.github/configs/amd-master.yaml b/.github/configs/amd-master.yaml
index bb433d6b1..61cdd6fe5 100644
--- a/.github/configs/amd-master.yaml
+++ b/.github/configs/amd-master.yaml
@@ -200,7 +200,7 @@ qwen3.5-bf16-mi325x-sglang:
       - { tp: 8, conc-start: 4, conc-end: 64 }
 
 qwen3.5-fp8-mi325x-sglang:
-  image: lmsysorg/sglang:v0.5.10-rocm720-mi30x
+  image: lmsysorg/sglang:v0.5.12-rocm720-mi30x
   model: Qwen/Qwen3.5-397B-A17B-FP8
   model-prefix: qwen3.5
   runner: mi325x
diff --git a/perf-changelog.yaml b/perf-changelog.yaml
index 4239ca63e..a8d5b259a 100644
--- a/perf-changelog.yaml
+++ b/perf-changelog.yaml
@@ -2586,3 +2586,9 @@
   description:
     - "Update vLLM image from v0.20.2 to v0.21.0"
   pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/1398
+
+- config-keys:
+    - qwen3.5-fp8-mi325x-sglang
+  description:
+    - "Update SGLang image from v0.5.10-rocm720-mi30x to v0.5.12-rocm720-mi30x"
+  pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/1430