diff --git a/configs/amd-master.yaml b/configs/amd-master.yaml
index 59f655945..84f5b24a1 100644
--- a/configs/amd-master.yaml
+++ b/configs/amd-master.yaml
@@ -769,7 +769,7 @@ dsr1-fp4-mi355x-sglang-disagg:
           - "DECODE_MTP_SIZE=0"
 
 dsr1-fp4-mi355x-sglang-disagg-8k1k-mtp:
-  image: lmsysorg/sglang-rocm:v0.5.12.post1-rocm720-mi35x-20260529
+  image: lmsysorg/sglang-rocm:v0.5.17-rocm720-mi35x-20260820
   model: amd/DeepSeek-R1-0528-MXFP4-v2
   model-prefix: dsr1
   runner: mi355x-disagg
diff --git a/perf-changelog.yaml b/perf-changelog.yaml
index 93b6a7c1b..1674f1a58 100644
--- a/perf-changelog.yaml
+++ b/perf-changelog.yaml
@@ -6251,3 +6251,10 @@
   description:
     - "Refresh with lower stream interval to collect correct client metrics"
   pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/2686
+
+- config-keys:
+    - dsr1-fp4-mi355x-sglang-disagg-8k1k-mtp
+  description:
+    - "Bump the SGLang ROCm image from v0.5.12.post1-rocm720-mi35x-20260529 to v0.5.17-rocm720-mi35x-20260820."
+    - "Pick up AITER d9e5ef7, including the stage-2 masked-reduce and persistent-MLA synchronization fixes for the DeepSeek-R1 MI355X accuracy corruption tracked in sgl-project/sglang#27194."
+  pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/2691