diff --git a/benchmarks/multi_node/amd_utils/job.slurm b/benchmarks/multi_node/amd_utils/job.slurm
index 895f1ac2fb..977722bf13 100755
--- a/benchmarks/multi_node/amd_utils/job.slurm
+++ b/benchmarks/multi_node/amd_utils/job.slurm
@@ -317,7 +317,7 @@ export DOCKER_CONT_NAME="container_${ENGINE}_${SANITIZED_USER}_${MODEL_NAME}_${S
 # vLLM external router container.
 # NOTE: vllm/vllm-router only retains ~16 recent nightlies on Docker Hub; older
 # dated tags are garbage-collected (manifest unknown)
-VLLM_ROUTER_IMAGE="${VLLM_ROUTER_IMAGE:-vllm/vllm-router:nightly-20260617-e667ebb}"
+VLLM_ROUTER_IMAGE="${VLLM_ROUTER_IMAGE:-vllm/vllm-router:nightly-20260629-e667ebb}"
 ROUTER_CONT_NAME="router_vllm_${SANITIZED_USER}_${SLURM_JOB_ID}"
 export RUN_FILE_FULL="$WS_PATH/${RUN_FILE}"
 
@@ -532,6 +532,7 @@ if [[ \"$ENGINE\" == \"vllm-disagg\" && \"$ROUTER_TYPE\" == \"vllm-router\" && \
     \$DOCKER_CMD run -d \
         --name \"$ROUTER_CONT_NAME\" \
         --network host \
+        --ulimit nofile=1048576:1048576 \
         -v /tmp:/run_logs \
         \"$VLLM_ROUTER_IMAGE\" \
         bash -lc \"mkdir -p /run_logs/slurm_job-${SLURM_JOB_ID} && exec vllm-router \
diff --git a/configs/amd-master.yaml b/configs/amd-master.yaml
index 72514a7756..d6b46cbbc0 100644
--- a/configs/amd-master.yaml
+++ b/configs/amd-master.yaml
@@ -1177,7 +1177,7 @@ dsr1-fp8-mi355x-sglang-disagg-mtp:
           - "DECODE_MTP_SIZE=2"
 
 kimik2.5-fp4-mi355x-vllm-disagg:
-  image: vllm/vllm-openai-rocm:nightly-bf610c2f56764e1b30bc6065f4ceace3d6e59036
+  image: vllm/vllm-openai-rocm:v0.24.0
   model: amd/Kimi-K2.5-MXFP4
   model-prefix: kimik2.5
   runner: mi355x-disagg
@@ -1200,7 +1200,6 @@ kimik2.5-fp4-mi355x-vllm-disagg:
           dp-attn: false
           additional-settings:
           - "PREFILL_NODES=1"
-          - "VLLM_MORIIO_CONNECTOR_READ_MODE=1"
         decode:
           num-worker: 2
           tp: 8
@@ -1221,7 +1220,6 @@ kimik2.5-fp4-mi355x-vllm-disagg:
           dp-attn: false
           additional-settings:
           - "PREFILL_NODES=1"
-          - "VLLM_MORIIO_CONNECTOR_READ_MODE=1"
         decode:
           num-worker: 2
           tp: 8
diff --git a/perf-changelog.yaml b/perf-changelog.yaml
index cd19882fc8..2b0b1acf39 100644
--- a/perf-changelog.yaml
+++ b/perf-changelog.yaml
@@ -4613,6 +4613,13 @@
     - "1k1k: 1 wide-EP max-throughput + 6 low-latency (DEP16/DEP32 + TP=4) topologies."
   pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/1780
 
+- config-keys:
+    - kimik2.5-fp4-mi355x-vllm-disagg
+  description:
+    - "Bump image to vllm/vllm-openai-rocm:v0.24.0 (MoRI-IO fixes upstreamed in vllm#40344)"
+    - "Rename a2a backend mori to mori_low_latency; set read_mode via kv_connector_extra_config instead of VLLM_MORIIO_CONNECTOR_READ_MODE env var"
+  pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/1585
+
 - config-keys:
     - dsr1-fp4-b200-dynamo-sglang-mtp
   description: