diff --git a/benchmarks/benchmark_lib.sh b/benchmarks/benchmark_lib.sh
index a4fe70803f..9cec822cca 100644
--- a/benchmarks/benchmark_lib.sh
+++ b/benchmarks/benchmark_lib.sh
@@ -1996,7 +1996,7 @@ build_replay_cmd() {
     # rolling TTFT/ITL/throughput block and emit it every 30 seconds.
     export AIPERF_UI_REALTIME_METRICS_ENABLED=true
     REPLAY_CMD="$AIPERF_CLI profile --scenario inferencex-agentx-mvp"
-    REPLAY_CMD+=" --url http://localhost:$PORT"
+    REPLAY_CMD+=" --url ${AIPERF_SERVER_URL:-http://localhost:$PORT}"
     REPLAY_CMD+=" --endpoint /v1/chat/completions"
     REPLAY_CMD+=" --endpoint-type chat"
     REPLAY_CMD+=" --streaming"
diff --git a/benchmarks/multi_node/agentic_srt.sh b/benchmarks/multi_node/agentic_srt.sh
index 79a36da524..77c196c133 100644
--- a/benchmarks/multi_node/agentic_srt.sh
+++ b/benchmarks/multi_node/agentic_srt.sh
@@ -9,7 +9,12 @@ set -x
 INFMAX_CONTAINER_WORKSPACE="${INFMAX_CONTAINER_WORKSPACE:-/infmax-workspace}"
 source "$INFMAX_CONTAINER_WORKSPACE/benchmarks/benchmark_lib.sh"
 
-check_env_vars MODEL MODEL_PREFIX FRAMEWORK PRECISION CONC RESULT_FILENAME DURATION
+check_env_vars \
+    MODEL MODEL_PREFIX FRAMEWORK PRECISION CONC RESULT_FILENAME DURATION \
+    SRT_FRONTEND_HOST SRT_FRONTEND_PORT
+
+export AIPERF_SERVER_URL="http://${SRT_FRONTEND_HOST}:${SRT_FRONTEND_PORT}"
+echo "Using srt-slurm frontend endpoint: $AIPERF_SERVER_URL"
 
 BASE_RESULT_DIR="${RESULT_DIR:-/logs/agentic}"
 BASE_RESULT_FILENAME="$RESULT_FILENAME"
@@ -32,7 +37,7 @@ install_agentic_deps
 wait_for_agentic_servers_idle() {
     local timeout_seconds="${AIPERF_DRAIN_TIMEOUT_SECONDS:-1800}"
     local poll_seconds="${AIPERF_DRAIN_POLL_SECONDS:-10}"
-    local frontend_metrics_url="http://localhost:${PORT}/metrics"
+    local frontend_metrics_url="${AIPERF_SERVER_URL%/}/metrics"
 
     "$AIPERF_PYTHON" - \
         "$timeout_seconds" \
diff --git a/benchmarks/multi_node/srt-slurm-recipes/trtllm/glm5.2/gb300-fp4/agentic/dynamo-agg-gb300-tp8-c1-b2-mtp8.yaml b/benchmarks/multi_node/srt-slurm-recipes/trtllm/glm5.2/gb300-fp4/agentic/dynamo-agg-gb300-tp8-c1-b2-mtp8.yaml
new file mode 100644
index 0000000000..eb3e3d2889
--- /dev/null
+++ b/benchmarks/multi_node/srt-slurm-recipes/trtllm/glm5.2/gb300-fp4/agentic/dynamo-agg-gb300-tp8-c1-b2-mtp8.yaml
@@ -0,0 +1,132 @@
+name: dynamo-agg-gb300-tp8-c1-b2-mtp8
+
+model:
+  path: nvidia/GLM-5.2-NVFP4
+  container: nvcr.io/nvidia/tensorrt-llm/release:1.3.0rc22.post1
+  precision: fp4
+
+dynamo:
+  install: true
+  wheel: 1.4.0.dev20260807
+  request_plane: tcp
+
+health_check:
+  max_attempts: 270
+  interval_seconds: 10
+
+resources:
+  gpu_type: gb300
+  gpus_per_node: 4
+  agg_nodes: 2
+  agg_workers: 1
+  gpus_per_agg: 8
+
+backend:
+  type: trtllm
+  aggregated_environment:
+    HF_HUB_OFFLINE: "1"
+    TRANSFORMERS_OFFLINE: "1"
+    TQDM_DISABLE: "1"
+    HF_HUB_DISABLE_PROGRESS_BARS: "1"
+    TLLM_LOG_LEVEL: INFO
+    TRTLLM_SERVER_DISABLE_GC: "1"
+    TRTLLM_WORKER_DISABLE_GC: "1"
+    TRTLLM_ENABLE_PDL: "1"
+    NCCL_GRAPH_MIXING_SUPPORT: "0"
+    MIMALLOC_PURGE_DELAY: "0"
+    PYTORCH_CUDA_ALLOC_CONF: expandable_segments:True
+    TRTLLM_SERVE_ENABLE_MSGSPEC: "1"
+    TLLM_ADP_ROUTER_MATCH_RATE_THRESHOLD: "0.10"
+    UCX_CUDA_IPC_ENABLE_MNNVL: "y"
+    UCX_RNDV_SCHEME: put_zcopy
+    TLLM_SPEC_DECODE_FORCE_NUM_ACCEPTED_TOKENS: "3.06"
+    TRTLLM_DISAGG_IDLE_CHECK: pr17324
+    DYN_TRTLLM_ENABLE_ATTENTION_DP: "1"
+    DYN_ENGINE_CONV_AFFINITY: "1"
+    DYN_TOKENIZER: fastokens
+    DYN_PUBLISH_KV_EVENTS: "0"
+  trtllm_config:
+    aggregated:
+      cuda_graph_config:
+        batch_sizes: [1, 2]
+        enable_padding: false
+      disable_overlap_scheduler: false
+      enable_attention_dp: false
+      enable_chunked_prefill: true
+      enable_lm_head_tp_in_adp: false
+      trust_remote_code: true
+      kv_cache_config:
+        dtype: fp8
+        enable_block_reuse: true
+        event_buffer_max_size: 0
+        free_gpu_memory_fraction: 0.85
+        tokens_per_block: 64
+      max_batch_size: 2
+      max_num_tokens: 8192
+      max_seq_len: 1048576
+      moe_config:
+        backend: TRTLLM
+        use_low_precision_moe_combine: true
+      moe_expert_parallel_size: 1
+      num_postprocess_workers: 8
+      perf_metrics_max_requests: 100000
+      pipeline_parallel_size: 1
+      print_iter_log: true
+      return_perf_metrics: false
+      sparse_attention_config:
+        algorithm: dsa
+        enable_heuristic_topk: true
+        use_cute_dsl_paged_mqa_logits: true
+      speculative_config:
+        decoding_type: MTP
+        max_draft_len: 8
+        num_nextn_predict_layers: 8
+      stream_interval: 20
+      tensor_parallel_size: 8
+  publish_events_and_metrics: false
+
+frontend:
+  type: dynamo
+  enable_multiple_frontends: false
+  env:
+    ETCD_LEASE_TTL: '120'
+    DYN_ROUTER_QUEUE_THRESHOLD: None
+    DYN_ROUTER_TEMPERATURE: '0'
+    DYN_TOKENIZER_CACHE: '1'
+    DYN_TOKENIZER_CACHE_BYTES: '8000000000'
+    DYN_TOKENIZER: fastokens
+    DYN_TCP_REQUEST_TIMEOUT: '30'
+    DYN_LOG: warn
+    DYN_ROUTER_SESSION_AFFINITY_TTL_SECS: '14400'
+  args:
+    router-mode: kv
+    no-kv-events: true
+    active-decode-blocks-threshold: None
+    active-prefill-tokens-threshold: None
+    active-prefill-tokens-threshold-frac: None
+
+benchmark:
+  type: custom
+  command: bash /infmax-workspace/benchmarks/multi_node/agentic_srt.sh
+  env:
+    INFMAX_CONTAINER_WORKSPACE: /infmax-workspace
+    RESULT_DIR: /logs/agentic
+    PORT: '8000'
+    IS_MULTINODE: 'true'
+    AIPERF_DYNAMO_SESSION_TIMEOUT_SECONDS: '14400'
+    AIPERF_DATASET_MMAP_CACHE_DIR: /aiperf_mmap_cache
+    HF_HUB_CACHE: /hf_hub_cache
+    WEKA_LOADER_OVERRIDE: semianalysis_cc_traces_weka_062126
+    AIPERF_DATASET_WEKA_LIVE_ASSISTANT_RESPONSES: "0"
+    SERVED_MODEL_NAME: GLM-5.2-NVFP4
+    MAX_MODEL_LEN: '1048576'
+    AIPERF_HTTP_X_DYNAMO_SESSION_ID_FROM_CORRELATION_ID: 'true'
+    AIPERF_USE_DYNAMO_CONV_AWARE_ROUTING: '0'
+    OPENAI_API_KEY: EMPTY
+    KV_OFFLOADING: none
+    MODEL: nvidia/GLM-5.2-NVFP4
+    MODEL_PREFIX: glm5.2
+    FRAMEWORK: dynamo-trt
+    PRECISION: fp4
+    CONC: '1'
+    DURATION: '3600'
diff --git a/benchmarks/multi_node/srt-slurm-recipes/trtllm/glm5.2/gb300-fp4/agentic/dynamo-disagg-gb300-1p1d-tep8-c20-b5-mtp5.yaml b/benchmarks/multi_node/srt-slurm-recipes/trtllm/glm5.2/gb300-fp4/agentic/dynamo-disagg-gb300-1p1d-tep8-c20-b5-mtp5.yaml
new file mode 100644
index 0000000000..033048cf9b
--- /dev/null
+++ b/benchmarks/multi_node/srt-slurm-recipes/trtllm/glm5.2/gb300-fp4/agentic/dynamo-disagg-gb300-1p1d-tep8-c20-b5-mtp5.yaml
@@ -0,0 +1,178 @@
+name: dynamo-disagg-gb300-1p1d-tep8-c20-b5-mtp5
+
+model:
+  path: nvidia/GLM-5.2-NVFP4
+  container: nvcr.io/nvidia/tensorrt-llm/release:1.3.0rc22.post1
+  precision: fp4
+
+dynamo:
+  install: true
+  wheel: 1.4.0.dev20260807
+  request_plane: tcp
+
+health_check:
+  max_attempts: 270
+  interval_seconds: 10
+
+resources:
+  gpu_type: gb300
+  gpus_per_node: 4
+  prefill_nodes: 1
+  prefill_workers: 1
+  gpus_per_prefill: 4
+  decode_nodes: 2
+  decode_workers: 1
+  gpus_per_decode: 8
+
+backend:
+  type: trtllm
+  prefill_environment: &server_environment
+    HF_HUB_OFFLINE: "1"
+    TRANSFORMERS_OFFLINE: "1"
+    TQDM_DISABLE: "1"
+    HF_HUB_DISABLE_PROGRESS_BARS: "1"
+    TLLM_LOG_LEVEL: INFO
+    TRTLLM_SERVER_DISABLE_GC: "1"
+    TRTLLM_WORKER_DISABLE_GC: "1"
+    TRTLLM_ENABLE_PDL: "1"
+    NCCL_GRAPH_MIXING_SUPPORT: "0"
+    MIMALLOC_PURGE_DELAY: "0"
+    PYTORCH_CUDA_ALLOC_CONF: expandable_segments:True
+    TRTLLM_SERVE_ENABLE_MSGSPEC: "1"
+    TLLM_ADP_ROUTER_MATCH_RATE_THRESHOLD: "0.10"
+    TRTLLM_KV_CACHE_TRANSFER_TIMEOUT_SEC: "600"
+    UCX_CUDA_IPC_ENABLE_MNNVL: "y"
+    UCX_RNDV_SCHEME: put_zcopy
+    TRTLLM_KVCACHE_SEND_BUFFER_COUNT: "1"
+    TRTLLM_KVCACHE_RECV_BUFFER_COUNT: "1"
+    TLLM_SPEC_DECODE_FORCE_NUM_ACCEPTED_TOKENS: "2.61"
+    DYN_TRTLLM_ENABLE_ATTENTION_DP: "1"
+    DYN_ENGINE_CONV_AFFINITY: "1"
+    DYN_TOKENIZER: fastokens
+    DYN_PUBLISH_KV_EVENTS: "0"
+  decode_environment: *server_environment
+  trtllm_config:
+    prefill:
+      attention_dp_config:
+        enable_kv_cache_aware_routing: false
+        kv_cache_routing_conversation_affinity: true
+        kv_cache_routing_max_sessions: 65536
+      cache_transceiver_config:
+        backend: NIXL
+        transceiver_runtime: PYTHON
+        kv_cache_bounce_size_mb: 5120
+        max_tokens_in_buffer: 1048576
+        kv_transfer_timeout_ms: 600000
+      cuda_graph_config: null
+      disable_overlap_scheduler: false
+      enable_attention_dp: true
+      enable_chunked_prefill: true
+      trust_remote_code: true
+      kv_cache_config:
+        dtype: fp8
+        enable_block_reuse: true
+        event_buffer_max_size: 0
+        free_gpu_memory_fraction: 0.75
+        host_cache_size: 137438953472
+        tokens_per_block: 64
+      max_batch_size: 256
+      max_num_tokens: 8192
+      max_seq_len: 1048576
+      moe_config:
+        backend: CUTEDSL
+      moe_expert_parallel_size: 4
+      num_postprocess_workers: 8
+      pipeline_parallel_size: 1
+      print_iter_log: true
+      return_perf_metrics: false
+      scheduler_config:
+        capacity_scheduler_policy: MAX_UTILIZATION
+      sparse_attention_config:
+        algorithm: dsa
+        enable_heuristic_topk: true
+      speculative_config:
+        decoding_type: MTP
+        max_draft_len: 5
+      tensor_parallel_size: 4
+    decode:
+      cache_transceiver_config:
+        backend: NIXL
+        transceiver_runtime: PYTHON
+        kv_cache_bounce_size_mb: 5120
+        max_tokens_in_buffer: 1048576
+        kv_transfer_timeout_ms: 600000
+      cuda_graph_config:
+        batch_sizes: [1, 2, 4, 5]
+        enable_padding: true
+      enable_attention_dp: false
+      trust_remote_code: true
+      kv_cache_config:
+        dtype: fp8
+        enable_block_reuse: false
+        event_buffer_max_size: 0
+        free_gpu_memory_fraction: 0.8
+        tokens_per_block: 64
+      max_batch_size: 5
+      max_num_tokens: 128
+      max_seq_len: 1048576
+      moe_config:
+        backend: TRTLLM
+        use_low_precision_moe_combine: true
+      moe_expert_parallel_size: 8
+      num_postprocess_workers: 4
+      pipeline_parallel_size: 1
+      print_iter_log: true
+      return_perf_metrics: false
+      sparse_attention_config:
+        algorithm: dsa
+        enable_heuristic_topk: true
+        use_cute_dsl_paged_mqa_logits: true
+        use_cute_dsl_topk: true
+      speculative_config:
+        decoding_type: MTP
+        max_draft_len: 5
+      stream_interval: 20
+      tensor_parallel_size: 8
+  publish_events_and_metrics: false
+
+frontend:
+  type: dynamo
+  enable_multiple_frontends: false
+  env:
+    ETCD_LEASE_TTL: '120'
+    DYN_ROUTER_QUEUE_THRESHOLD: None
+    DYN_ROUTER_TEMPERATURE: '0'
+    DYN_TOKENIZER_CACHE: '1'
+    DYN_TOKENIZER_CACHE_BYTES: '8000000000'
+    DYN_TOKENIZER: fastokens
+    DYN_TCP_REQUEST_TIMEOUT: '30'
+    DYN_LOG: warn
+    DYN_ROUTER_SESSION_AFFINITY_TTL_SECS: '14400'
+  args:
+    router-mode: kv
+    no-kv-events: true
+    active-decode-blocks-threshold: None
+    active-prefill-tokens-threshold: None
+    active-prefill-tokens-threshold-frac: None
+  orchestrator_placement: head
+
+benchmark:
+  type: custom
+  client_placement: last_decode
+  command: bash /infmax-workspace/benchmarks/multi_node/agentic_srt.sh
+  env:
+    INFMAX_CONTAINER_WORKSPACE: /infmax-workspace
+    RESULT_DIR: /logs/agentic
+    PORT: '8000'
+    IS_MULTINODE: 'true'
+    AIPERF_DYNAMO_SESSION_TIMEOUT_SECONDS: '14400'
+    AIPERF_DATASET_MMAP_CACHE_DIR: /aiperf_mmap_cache
+    HF_HUB_CACHE: /hf_hub_cache
+    WEKA_LOADER_OVERRIDE: semianalysis_cc_traces_weka_062126
+    AIPERF_DATASET_WEKA_LIVE_ASSISTANT_RESPONSES: "0"
+    SERVED_MODEL_NAME: GLM-5.2-NVFP4
+    MAX_MODEL_LEN: '1048576'
+    AIPERF_HTTP_X_DYNAMO_SESSION_ID_FROM_CORRELATION_ID: 'true'
+    AIPERF_USE_DYNAMO_CONV_AWARE_ROUTING: '0'
+    OPENAI_API_KEY: EMPTY
+    KV_OFFLOADING: none
diff --git a/benchmarks/multi_node/srt-slurm-recipes/trtllm/glm5.2/gb300-fp4/agentic/dynamo-disagg-gb300-1p4d-tep4-c30-b2-mtp5.yaml b/benchmarks/multi_node/srt-slurm-recipes/trtllm/glm5.2/gb300-fp4/agentic/dynamo-disagg-gb300-1p4d-tep4-c30-b2-mtp5.yaml
new file mode 100644
index 0000000000..8f6fdcc2cf
--- /dev/null
+++ b/benchmarks/multi_node/srt-slurm-recipes/trtllm/glm5.2/gb300-fp4/agentic/dynamo-disagg-gb300-1p4d-tep4-c30-b2-mtp5.yaml
@@ -0,0 +1,178 @@
+name: dynamo-disagg-gb300-1p4d-tep4-c30-b2-mtp5
+
+model:
+  path: nvidia/GLM-5.2-NVFP4
+  container: nvcr.io/nvidia/tensorrt-llm/release:1.3.0rc22.post1
+  precision: fp4
+
+dynamo:
+  install: true
+  wheel: 1.4.0.dev20260807
+  request_plane: tcp
+
+health_check:
+  max_attempts: 270
+  interval_seconds: 10
+
+resources:
+  gpu_type: gb300
+  gpus_per_node: 4
+  prefill_nodes: 1
+  prefill_workers: 1
+  gpus_per_prefill: 4
+  decode_nodes: 4
+  decode_workers: 4
+  gpus_per_decode: 4
+
+backend:
+  type: trtllm
+  prefill_environment: &server_environment
+    HF_HUB_OFFLINE: "1"
+    TRANSFORMERS_OFFLINE: "1"
+    TQDM_DISABLE: "1"
+    HF_HUB_DISABLE_PROGRESS_BARS: "1"
+    TLLM_LOG_LEVEL: INFO
+    TRTLLM_SERVER_DISABLE_GC: "1"
+    TRTLLM_WORKER_DISABLE_GC: "1"
+    TRTLLM_ENABLE_PDL: "1"
+    NCCL_GRAPH_MIXING_SUPPORT: "0"
+    MIMALLOC_PURGE_DELAY: "0"
+    PYTORCH_CUDA_ALLOC_CONF: expandable_segments:True
+    TRTLLM_SERVE_ENABLE_MSGSPEC: "1"
+    TLLM_ADP_ROUTER_MATCH_RATE_THRESHOLD: "0.10"
+    TRTLLM_KV_CACHE_TRANSFER_TIMEOUT_SEC: "600"
+    UCX_CUDA_IPC_ENABLE_MNNVL: "y"
+    UCX_RNDV_SCHEME: put_zcopy
+    TRTLLM_KVCACHE_SEND_BUFFER_COUNT: "1"
+    TRTLLM_KVCACHE_RECV_BUFFER_COUNT: "1"
+    TLLM_SPEC_DECODE_FORCE_NUM_ACCEPTED_TOKENS: "2.61"
+    DYN_TRTLLM_ENABLE_ATTENTION_DP: "1"
+    DYN_ENGINE_CONV_AFFINITY: "1"
+    DYN_TOKENIZER: fastokens
+    DYN_PUBLISH_KV_EVENTS: "0"
+  decode_environment: *server_environment
+  trtllm_config:
+    prefill:
+      attention_dp_config:
+        enable_kv_cache_aware_routing: false
+        kv_cache_routing_conversation_affinity: true
+        kv_cache_routing_max_sessions: 65536
+      cache_transceiver_config:
+        backend: NIXL
+        transceiver_runtime: PYTHON
+        kv_cache_bounce_size_mb: 5120
+        max_tokens_in_buffer: 1048576
+        kv_transfer_timeout_ms: 600000
+      cuda_graph_config: null
+      disable_overlap_scheduler: false
+      enable_attention_dp: true
+      enable_chunked_prefill: true
+      trust_remote_code: true
+      kv_cache_config:
+        dtype: fp8
+        enable_block_reuse: true
+        event_buffer_max_size: 0
+        free_gpu_memory_fraction: 0.75
+        host_cache_size: 137438953472
+        tokens_per_block: 64
+      max_batch_size: 256
+      max_num_tokens: 8192
+      max_seq_len: 1048576
+      moe_config:
+        backend: CUTEDSL
+      moe_expert_parallel_size: 4
+      num_postprocess_workers: 8
+      pipeline_parallel_size: 1
+      print_iter_log: true
+      return_perf_metrics: false
+      scheduler_config:
+        capacity_scheduler_policy: MAX_UTILIZATION
+      sparse_attention_config:
+        algorithm: dsa
+        enable_heuristic_topk: true
+      speculative_config:
+        decoding_type: MTP
+        max_draft_len: 5
+      tensor_parallel_size: 4
+    decode:
+      cache_transceiver_config:
+        backend: NIXL
+        transceiver_runtime: PYTHON
+        kv_cache_bounce_size_mb: 5120
+        max_tokens_in_buffer: 1048576
+        kv_transfer_timeout_ms: 600000
+      cuda_graph_config:
+        batch_sizes: [1, 2]
+        enable_padding: true
+      enable_attention_dp: false
+      trust_remote_code: true
+      kv_cache_config:
+        dtype: fp8
+        enable_block_reuse: false
+        event_buffer_max_size: 0
+        free_gpu_memory_fraction: 0.8
+        tokens_per_block: 64
+      max_batch_size: 2
+      max_num_tokens: 128
+      max_seq_len: 1048576
+      moe_config:
+        backend: TRTLLM
+        use_low_precision_moe_combine: true
+      moe_expert_parallel_size: 4
+      num_postprocess_workers: 4
+      pipeline_parallel_size: 1
+      print_iter_log: true
+      return_perf_metrics: false
+      sparse_attention_config:
+        algorithm: dsa
+        enable_heuristic_topk: true
+        use_cute_dsl_paged_mqa_logits: true
+        use_cute_dsl_topk: true
+      speculative_config:
+        decoding_type: MTP
+        max_draft_len: 5
+      stream_interval: 20
+      tensor_parallel_size: 4
+  publish_events_and_metrics: false
+
+frontend:
+  type: dynamo
+  enable_multiple_frontends: false
+  env:
+    ETCD_LEASE_TTL: '120'
+    DYN_ROUTER_QUEUE_THRESHOLD: None
+    DYN_ROUTER_TEMPERATURE: '0'
+    DYN_TOKENIZER_CACHE: '1'
+    DYN_TOKENIZER_CACHE_BYTES: '8000000000'
+    DYN_TOKENIZER: fastokens
+    DYN_TCP_REQUEST_TIMEOUT: '30'
+    DYN_LOG: warn
+    DYN_ROUTER_SESSION_AFFINITY_TTL_SECS: '14400'
+  args:
+    router-mode: kv
+    no-kv-events: true
+    active-decode-blocks-threshold: None
+    active-prefill-tokens-threshold: None
+    active-prefill-tokens-threshold-frac: None
+  orchestrator_placement: head
+
+benchmark:
+  type: custom
+  client_placement: last_decode
+  command: bash /infmax-workspace/benchmarks/multi_node/agentic_srt.sh
+  env:
+    INFMAX_CONTAINER_WORKSPACE: /infmax-workspace
+    RESULT_DIR: /logs/agentic
+    PORT: '8000'
+    IS_MULTINODE: 'true'
+    AIPERF_DYNAMO_SESSION_TIMEOUT_SECONDS: '14400'
+    AIPERF_DATASET_MMAP_CACHE_DIR: /aiperf_mmap_cache
+    HF_HUB_CACHE: /hf_hub_cache
+    WEKA_LOADER_OVERRIDE: semianalysis_cc_traces_weka_062126
+    AIPERF_DATASET_WEKA_LIVE_ASSISTANT_RESPONSES: "0"
+    SERVED_MODEL_NAME: GLM-5.2-NVFP4
+    MAX_MODEL_LEN: '1048576'
+    AIPERF_HTTP_X_DYNAMO_SESSION_ID_FROM_CORRELATION_ID: 'true'
+    AIPERF_USE_DYNAMO_CONV_AWARE_ROUTING: '0'
+    OPENAI_API_KEY: EMPTY
+    KV_OFFLOADING: none
diff --git a/benchmarks/multi_node/srt-slurm-recipes/trtllm/glm5.2/gb300-fp4/agentic/dynamo-disagg-gb300-3p4d-tep4-c60-b5-mtp5.yaml b/benchmarks/multi_node/srt-slurm-recipes/trtllm/glm5.2/gb300-fp4/agentic/dynamo-disagg-gb300-3p4d-tep4-c60-b5-mtp5.yaml
new file mode 100644
index 0000000000..c618bf7f89
--- /dev/null
+++ b/benchmarks/multi_node/srt-slurm-recipes/trtllm/glm5.2/gb300-fp4/agentic/dynamo-disagg-gb300-3p4d-tep4-c60-b5-mtp5.yaml
@@ -0,0 +1,178 @@
+name: dynamo-disagg-gb300-3p4d-tep4-c60-b5-mtp5
+
+model:
+  path: nvidia/GLM-5.2-NVFP4
+  container: nvcr.io/nvidia/tensorrt-llm/release:1.3.0rc22.post1
+  precision: fp4
+
+dynamo:
+  install: true
+  wheel: 1.4.0.dev20260807
+  request_plane: tcp
+
+health_check:
+  max_attempts: 270
+  interval_seconds: 10
+
+resources:
+  gpu_type: gb300
+  gpus_per_node: 4
+  prefill_nodes: 3
+  prefill_workers: 3
+  gpus_per_prefill: 4
+  decode_nodes: 4
+  decode_workers: 4
+  gpus_per_decode: 4
+
+backend:
+  type: trtllm
+  prefill_environment: &server_environment
+    HF_HUB_OFFLINE: "1"
+    TRANSFORMERS_OFFLINE: "1"
+    TQDM_DISABLE: "1"
+    HF_HUB_DISABLE_PROGRESS_BARS: "1"
+    TLLM_LOG_LEVEL: INFO
+    TRTLLM_SERVER_DISABLE_GC: "1"
+    TRTLLM_WORKER_DISABLE_GC: "1"
+    TRTLLM_ENABLE_PDL: "1"
+    NCCL_GRAPH_MIXING_SUPPORT: "0"
+    MIMALLOC_PURGE_DELAY: "0"
+    PYTORCH_CUDA_ALLOC_CONF: expandable_segments:True
+    TRTLLM_SERVE_ENABLE_MSGSPEC: "1"
+    TLLM_ADP_ROUTER_MATCH_RATE_THRESHOLD: "0.10"
+    TRTLLM_KV_CACHE_TRANSFER_TIMEOUT_SEC: "600"
+    UCX_CUDA_IPC_ENABLE_MNNVL: "y"
+    UCX_RNDV_SCHEME: put_zcopy
+    TRTLLM_KVCACHE_SEND_BUFFER_COUNT: "1"
+    TRTLLM_KVCACHE_RECV_BUFFER_COUNT: "1"
+    TLLM_SPEC_DECODE_FORCE_NUM_ACCEPTED_TOKENS: "2.61"
+    DYN_TRTLLM_ENABLE_ATTENTION_DP: "1"
+    DYN_ENGINE_CONV_AFFINITY: "1"
+    DYN_TOKENIZER: fastokens
+    DYN_PUBLISH_KV_EVENTS: "0"
+  decode_environment: *server_environment
+  trtllm_config:
+    prefill:
+      attention_dp_config:
+        enable_kv_cache_aware_routing: false
+        kv_cache_routing_conversation_affinity: true
+        kv_cache_routing_max_sessions: 65536
+      cache_transceiver_config:
+        backend: NIXL
+        transceiver_runtime: PYTHON
+        kv_cache_bounce_size_mb: 5120
+        max_tokens_in_buffer: 1048576
+        kv_transfer_timeout_ms: 600000
+      cuda_graph_config: null
+      disable_overlap_scheduler: false
+      enable_attention_dp: true
+      enable_chunked_prefill: true
+      trust_remote_code: true
+      kv_cache_config:
+        dtype: fp8
+        enable_block_reuse: true
+        event_buffer_max_size: 0
+        free_gpu_memory_fraction: 0.75
+        host_cache_size: 137438953472
+        tokens_per_block: 64
+      max_batch_size: 256
+      max_num_tokens: 8192
+      max_seq_len: 1048576
+      moe_config:
+        backend: CUTEDSL
+      moe_expert_parallel_size: 4
+      num_postprocess_workers: 8
+      pipeline_parallel_size: 1
+      print_iter_log: true
+      return_perf_metrics: false
+      scheduler_config:
+        capacity_scheduler_policy: MAX_UTILIZATION
+      sparse_attention_config:
+        algorithm: dsa
+        enable_heuristic_topk: true
+      speculative_config:
+        decoding_type: MTP
+        max_draft_len: 5
+      tensor_parallel_size: 4
+    decode:
+      cache_transceiver_config:
+        backend: NIXL
+        transceiver_runtime: PYTHON
+        kv_cache_bounce_size_mb: 5120
+        max_tokens_in_buffer: 1048576
+        kv_transfer_timeout_ms: 600000
+      cuda_graph_config:
+        batch_sizes: [1, 2, 4, 5]
+        enable_padding: true
+      enable_attention_dp: false
+      trust_remote_code: true
+      kv_cache_config:
+        dtype: fp8
+        enable_block_reuse: false
+        event_buffer_max_size: 0
+        free_gpu_memory_fraction: 0.8
+        tokens_per_block: 64
+      max_batch_size: 5
+      max_num_tokens: 128
+      max_seq_len: 1048576
+      moe_config:
+        backend: TRTLLM
+        use_low_precision_moe_combine: true
+      moe_expert_parallel_size: 4
+      num_postprocess_workers: 4
+      pipeline_parallel_size: 1
+      print_iter_log: true
+      return_perf_metrics: false
+      sparse_attention_config:
+        algorithm: dsa
+        enable_heuristic_topk: true
+        use_cute_dsl_paged_mqa_logits: true
+        use_cute_dsl_topk: true
+      speculative_config:
+        decoding_type: MTP
+        max_draft_len: 5
+      stream_interval: 20
+      tensor_parallel_size: 4
+  publish_events_and_metrics: false
+
+frontend:
+  type: dynamo
+  enable_multiple_frontends: false
+  env:
+    ETCD_LEASE_TTL: '120'
+    DYN_ROUTER_QUEUE_THRESHOLD: None
+    DYN_ROUTER_TEMPERATURE: '0'
+    DYN_TOKENIZER_CACHE: '1'
+    DYN_TOKENIZER_CACHE_BYTES: '8000000000'
+    DYN_TOKENIZER: fastokens
+    DYN_TCP_REQUEST_TIMEOUT: '30'
+    DYN_LOG: warn
+    DYN_ROUTER_SESSION_AFFINITY_TTL_SECS: '14400'
+  args:
+    router-mode: kv
+    no-kv-events: true
+    active-decode-blocks-threshold: None
+    active-prefill-tokens-threshold: None
+    active-prefill-tokens-threshold-frac: None
+  orchestrator_placement: head
+
+benchmark:
+  type: custom
+  client_placement: last_decode
+  command: bash /infmax-workspace/benchmarks/multi_node/agentic_srt.sh
+  env:
+    INFMAX_CONTAINER_WORKSPACE: /infmax-workspace
+    RESULT_DIR: /logs/agentic
+    PORT: '8000'
+    IS_MULTINODE: 'true'
+    AIPERF_DYNAMO_SESSION_TIMEOUT_SECONDS: '14400'
+    AIPERF_DATASET_MMAP_CACHE_DIR: /aiperf_mmap_cache
+    HF_HUB_CACHE: /hf_hub_cache
+    WEKA_LOADER_OVERRIDE: semianalysis_cc_traces_weka_062126
+    AIPERF_DATASET_WEKA_LIVE_ASSISTANT_RESPONSES: "0"
+    SERVED_MODEL_NAME: GLM-5.2-NVFP4
+    MAX_MODEL_LEN: '1048576'
+    AIPERF_HTTP_X_DYNAMO_SESSION_ID_FROM_CORRELATION_ID: 'true'
+    AIPERF_USE_DYNAMO_CONV_AWARE_ROUTING: '0'
+    OPENAI_API_KEY: EMPTY
+    KV_OFFLOADING: none
diff --git a/benchmarks/multi_node/srt-slurm-recipes/trtllm/glm5.2/gb300-fp4/agentic/dynamo-disagg-gb300-5p1d-dep16-c152-b16-mtp3.yaml b/benchmarks/multi_node/srt-slurm-recipes/trtllm/glm5.2/gb300-fp4/agentic/dynamo-disagg-gb300-5p1d-dep16-c152-b16-mtp3.yaml
new file mode 100644
index 0000000000..8407308128
--- /dev/null
+++ b/benchmarks/multi_node/srt-slurm-recipes/trtllm/glm5.2/gb300-fp4/agentic/dynamo-disagg-gb300-5p1d-dep16-c152-b16-mtp3.yaml
@@ -0,0 +1,179 @@
+name: dynamo-disagg-gb300-5p1d-dep16-c152-b16-mtp3
+
+model:
+  path: nvidia/GLM-5.2-NVFP4
+  container: nvcr.io/nvidia/tensorrt-llm/release:1.3.0rc22.post1
+  precision: fp4
+
+dynamo:
+  install: true
+  wheel: 1.4.0.dev20260807
+  request_plane: tcp
+
+health_check:
+  max_attempts: 270
+  interval_seconds: 10
+
+resources:
+  gpu_type: gb300
+  gpus_per_node: 4
+  prefill_nodes: 5
+  prefill_workers: 5
+  gpus_per_prefill: 4
+  decode_nodes: 4
+  decode_workers: 1
+  gpus_per_decode: 16
+
+backend:
+  type: trtllm
+  prefill_environment: &server_environment
+    HF_HUB_OFFLINE: "1"
+    TRANSFORMERS_OFFLINE: "1"
+    TQDM_DISABLE: "1"
+    HF_HUB_DISABLE_PROGRESS_BARS: "1"
+    TLLM_LOG_LEVEL: INFO
+    TRTLLM_SERVER_DISABLE_GC: "1"
+    TRTLLM_WORKER_DISABLE_GC: "1"
+    TRTLLM_ENABLE_PDL: "1"
+    NCCL_GRAPH_MIXING_SUPPORT: "0"
+    MIMALLOC_PURGE_DELAY: "0"
+    PYTORCH_CUDA_ALLOC_CONF: expandable_segments:True
+    TRTLLM_SERVE_ENABLE_MSGSPEC: "1"
+    TLLM_ADP_ROUTER_MATCH_RATE_THRESHOLD: "0.10"
+    TRTLLM_KV_CACHE_TRANSFER_TIMEOUT_SEC: "600"
+    UCX_CUDA_IPC_ENABLE_MNNVL: "y"
+    UCX_RNDV_SCHEME: put_zcopy
+    TRTLLM_KVCACHE_SEND_BUFFER_COUNT: "1"
+    TRTLLM_KVCACHE_RECV_BUFFER_COUNT: "1"
+    TLLM_SPEC_DECODE_FORCE_NUM_ACCEPTED_TOKENS: "1.99"
+    DYN_TRTLLM_ENABLE_ATTENTION_DP: "1"
+    DYN_ENGINE_CONV_AFFINITY: "1"
+    DYN_TOKENIZER: fastokens
+    DYN_PUBLISH_KV_EVENTS: "0"
+  decode_environment: *server_environment
+  trtllm_config:
+    prefill:
+      attention_dp_config:
+        enable_kv_cache_aware_routing: false
+        kv_cache_routing_conversation_affinity: true
+        kv_cache_routing_max_sessions: 65536
+      cache_transceiver_config:
+        backend: NIXL
+        transceiver_runtime: PYTHON
+        kv_cache_bounce_size_mb: 5120
+        max_tokens_in_buffer: 1048576
+        kv_transfer_timeout_ms: 600000
+      cuda_graph_config: null
+      disable_overlap_scheduler: false
+      enable_attention_dp: true
+      enable_chunked_prefill: true
+      trust_remote_code: true
+      kv_cache_config:
+        dtype: fp8
+        enable_block_reuse: true
+        event_buffer_max_size: 0
+        free_gpu_memory_fraction: 0.75
+        host_cache_size: 137438953472
+        tokens_per_block: 64
+      max_batch_size: 256
+      max_num_tokens: 8192
+      max_seq_len: 1048576
+      moe_config:
+        backend: CUTEDSL
+      moe_expert_parallel_size: 4
+      num_postprocess_workers: 8
+      pipeline_parallel_size: 1
+      print_iter_log: true
+      return_perf_metrics: false
+      scheduler_config:
+        capacity_scheduler_policy: MAX_UTILIZATION
+      sparse_attention_config:
+        algorithm: dsa
+        enable_heuristic_topk: true
+      speculative_config:
+        decoding_type: MTP
+        max_draft_len: 3
+      tensor_parallel_size: 4
+    decode:
+      cache_transceiver_config:
+        backend: NIXL
+        transceiver_runtime: PYTHON
+        kv_cache_bounce_size_mb: 5120
+        max_tokens_in_buffer: 1048576
+        kv_transfer_timeout_ms: 600000
+      cuda_graph_config:
+        batch_sizes: [1, 2, 4, 8, 16]
+        enable_padding: true
+      enable_attention_dp: true
+      enable_lm_head_tp_in_adp: false
+      trust_remote_code: true
+      kv_cache_config:
+        dtype: fp8
+        enable_block_reuse: false
+        event_buffer_max_size: 0
+        free_gpu_memory_fraction: 0.9
+        tokens_per_block: 64
+      max_batch_size: 16
+      max_num_tokens: 128
+      max_seq_len: 1048576
+      moe_config:
+        backend: TRTLLM
+        use_low_precision_moe_combine: true
+      moe_expert_parallel_size: 16
+      num_postprocess_workers: 4
+      pipeline_parallel_size: 1
+      print_iter_log: true
+      return_perf_metrics: false
+      sparse_attention_config:
+        algorithm: dsa
+        enable_heuristic_topk: true
+        use_cute_dsl_paged_mqa_logits: true
+        use_cute_dsl_topk: true
+      speculative_config:
+        decoding_type: MTP
+        max_draft_len: 3
+      stream_interval: 20
+      tensor_parallel_size: 16
+  publish_events_and_metrics: false
+
+frontend:
+  type: dynamo
+  enable_multiple_frontends: false
+  env:
+    ETCD_LEASE_TTL: '120'
+    DYN_ROUTER_QUEUE_THRESHOLD: None
+    DYN_ROUTER_TEMPERATURE: '0'
+    DYN_TOKENIZER_CACHE: '1'
+    DYN_TOKENIZER_CACHE_BYTES: '8000000000'
+    DYN_TOKENIZER: fastokens
+    DYN_TCP_REQUEST_TIMEOUT: '30'
+    DYN_LOG: warn
+    DYN_ROUTER_SESSION_AFFINITY_TTL_SECS: '14400'
+  args:
+    router-mode: kv
+    no-kv-events: true
+    active-decode-blocks-threshold: None
+    active-prefill-tokens-threshold: None
+    active-prefill-tokens-threshold-frac: None
+  orchestrator_placement: head
+
+benchmark:
+  type: custom
+  client_placement: last_decode
+  command: bash /infmax-workspace/benchmarks/multi_node/agentic_srt.sh
+  env:
+    INFMAX_CONTAINER_WORKSPACE: /infmax-workspace
+    RESULT_DIR: /logs/agentic
+    PORT: '8000'
+    IS_MULTINODE: 'true'
+    AIPERF_DYNAMO_SESSION_TIMEOUT_SECONDS: '14400'
+    AIPERF_DATASET_MMAP_CACHE_DIR: /aiperf_mmap_cache
+    HF_HUB_CACHE: /hf_hub_cache
+    WEKA_LOADER_OVERRIDE: semianalysis_cc_traces_weka_062126
+    AIPERF_DATASET_WEKA_LIVE_ASSISTANT_RESPONSES: "0"
+    SERVED_MODEL_NAME: GLM-5.2-NVFP4
+    MAX_MODEL_LEN: '1048576'
+    AIPERF_HTTP_X_DYNAMO_SESSION_ID_FROM_CORRELATION_ID: 'true'
+    AIPERF_USE_DYNAMO_CONV_AWARE_ROUTING: '0'
+    OPENAI_API_KEY: EMPTY
+    KV_OFFLOADING: none
diff --git a/benchmarks/multi_node/srt-slurm-recipes/trtllm/glm5.2/gb300-fp4/agentic/dynamo-disagg-gb300-6p1d-dep8-c227-b16-mtp3.yaml b/benchmarks/multi_node/srt-slurm-recipes/trtllm/glm5.2/gb300-fp4/agentic/dynamo-disagg-gb300-6p1d-dep8-c227-b16-mtp3.yaml
new file mode 100644
index 0000000000..4941ce4270
--- /dev/null
+++ b/benchmarks/multi_node/srt-slurm-recipes/trtllm/glm5.2/gb300-fp4/agentic/dynamo-disagg-gb300-6p1d-dep8-c227-b16-mtp3.yaml
@@ -0,0 +1,178 @@
+name: dynamo-disagg-gb300-6p1d-dep8-c227-b16-mtp3
+
+model:
+  path: nvidia/GLM-5.2-NVFP4
+  container: nvcr.io/nvidia/tensorrt-llm/release:1.3.0rc22.post1
+  precision: fp4
+
+dynamo:
+  install: true
+  wheel: 1.4.0.dev20260807
+  request_plane: tcp
+
+health_check:
+  max_attempts: 270
+  interval_seconds: 10
+
+resources:
+  gpu_type: gb300
+  gpus_per_node: 4
+  prefill_nodes: 6
+  prefill_workers: 6
+  gpus_per_prefill: 4
+  decode_nodes: 2
+  decode_workers: 1
+  gpus_per_decode: 8
+
+backend:
+  type: trtllm
+  prefill_environment: &server_environment
+    HF_HUB_OFFLINE: "1"
+    TRANSFORMERS_OFFLINE: "1"
+    TQDM_DISABLE: "1"
+    HF_HUB_DISABLE_PROGRESS_BARS: "1"
+    TLLM_LOG_LEVEL: INFO
+    TRTLLM_SERVER_DISABLE_GC: "1"
+    TRTLLM_WORKER_DISABLE_GC: "1"
+    TRTLLM_ENABLE_PDL: "1"
+    NCCL_GRAPH_MIXING_SUPPORT: "0"
+    MIMALLOC_PURGE_DELAY: "0"
+    PYTORCH_CUDA_ALLOC_CONF: expandable_segments:True
+    TRTLLM_SERVE_ENABLE_MSGSPEC: "1"
+    TLLM_ADP_ROUTER_MATCH_RATE_THRESHOLD: "0.10"
+    TRTLLM_KV_CACHE_TRANSFER_TIMEOUT_SEC: "600"
+    UCX_CUDA_IPC_ENABLE_MNNVL: "y"
+    UCX_RNDV_SCHEME: put_zcopy
+    TRTLLM_KVCACHE_SEND_BUFFER_COUNT: "1"
+    TRTLLM_KVCACHE_RECV_BUFFER_COUNT: "1"
+    TLLM_SPEC_DECODE_FORCE_NUM_ACCEPTED_TOKENS: "1.99"
+    DYN_TRTLLM_ENABLE_ATTENTION_DP: "1"
+    DYN_ENGINE_CONV_AFFINITY: "1"
+    DYN_TOKENIZER: fastokens
+    DYN_PUBLISH_KV_EVENTS: "0"
+  decode_environment: *server_environment
+  trtllm_config:
+    prefill:
+      attention_dp_config:
+        enable_kv_cache_aware_routing: false
+        kv_cache_routing_conversation_affinity: true
+        kv_cache_routing_max_sessions: 65536
+      cache_transceiver_config:
+        backend: NIXL
+        transceiver_runtime: PYTHON
+        kv_cache_bounce_size_mb: 5120
+        max_tokens_in_buffer: 1048576
+        kv_transfer_timeout_ms: 600000
+      cuda_graph_config: null
+      disable_overlap_scheduler: false
+      enable_attention_dp: true
+      enable_chunked_prefill: true
+      trust_remote_code: true
+      kv_cache_config:
+        dtype: fp8
+        enable_block_reuse: true
+        event_buffer_max_size: 0
+        free_gpu_memory_fraction: 0.75
+        host_cache_size: 137438953472
+        tokens_per_block: 64
+      max_batch_size: 256
+      max_num_tokens: 8192
+      max_seq_len: 1048576
+      moe_config:
+        backend: CUTEDSL
+      moe_expert_parallel_size: 4
+      num_postprocess_workers: 8
+      pipeline_parallel_size: 1
+      print_iter_log: true
+      return_perf_metrics: false
+      scheduler_config:
+        capacity_scheduler_policy: MAX_UTILIZATION
+      sparse_attention_config:
+        algorithm: dsa
+        enable_heuristic_topk: true
+      speculative_config:
+        decoding_type: MTP
+        max_draft_len: 3
+      tensor_parallel_size: 4
+    decode:
+      cache_transceiver_config:
+        backend: NIXL
+        transceiver_runtime: PYTHON
+        kv_cache_bounce_size_mb: 5120
+        max_tokens_in_buffer: 1048576
+        kv_transfer_timeout_ms: 600000
+      cuda_graph_config:
+        batch_sizes: [1, 2, 4, 8, 16]
+        enable_padding: true
+      enable_attention_dp: true
+      enable_lm_head_tp_in_adp: false
+      trust_remote_code: true
+      kv_cache_config:
+        dtype: fp8
+        enable_block_reuse: false
+        event_buffer_max_size: 0
+        free_gpu_memory_fraction: 0.9
+        tokens_per_block: 64
+      max_batch_size: 16
+      max_num_tokens: 128
+      max_seq_len: 1048576
+      moe_config:
+        backend: CUTEDSL
+        use_low_precision_moe_combine: true
+      moe_expert_parallel_size: 8
+      num_postprocess_workers: 4
+      pipeline_parallel_size: 1
+      print_iter_log: true
+      return_perf_metrics: false
+      sparse_attention_config:
+        algorithm: dsa
+        enable_heuristic_topk: true
+        use_cute_dsl_paged_mqa_logits: true
+      speculative_config:
+        decoding_type: MTP
+        max_draft_len: 3
+      stream_interval: 20
+      tensor_parallel_size: 8
+  publish_events_and_metrics: false
+
+frontend:
+  type: dynamo
+  enable_multiple_frontends: false
+  env:
+    ETCD_LEASE_TTL: '120'
+    DYN_ROUTER_QUEUE_THRESHOLD: None
+    DYN_ROUTER_TEMPERATURE: '0'
+    DYN_TOKENIZER_CACHE: '1'
+    DYN_TOKENIZER_CACHE_BYTES: '8000000000'
+    DYN_TOKENIZER: fastokens
+    DYN_TCP_REQUEST_TIMEOUT: '30'
+    DYN_LOG: warn
+    DYN_ROUTER_SESSION_AFFINITY_TTL_SECS: '14400'
+  args:
+    router-mode: kv
+    no-kv-events: true
+    active-decode-blocks-threshold: None
+    active-prefill-tokens-threshold: None
+    active-prefill-tokens-threshold-frac: None
+  orchestrator_placement: head
+
+benchmark:
+  type: custom
+  client_placement: last_decode
+  command: bash /infmax-workspace/benchmarks/multi_node/agentic_srt.sh
+  env:
+    INFMAX_CONTAINER_WORKSPACE: /infmax-workspace
+    RESULT_DIR: /logs/agentic
+    PORT: '8000'
+    IS_MULTINODE: 'true'
+    AIPERF_DYNAMO_SESSION_TIMEOUT_SECONDS: '14400'
+    AIPERF_DATASET_MMAP_CACHE_DIR: /aiperf_mmap_cache
+    HF_HUB_CACHE: /hf_hub_cache
+    WEKA_LOADER_OVERRIDE: semianalysis_cc_traces_weka_062126
+    AIPERF_DATASET_WEKA_LIVE_ASSISTANT_RESPONSES: "0"
+    SERVED_MODEL_NAME: GLM-5.2-NVFP4
+    MAX_MODEL_LEN: '1048576'
+    AIPERF_HTTP_X_DYNAMO_SESSION_ID_FROM_CORRELATION_ID: 'true'
+    AIPERF_USE_DYNAMO_CONV_AWARE_ROUTING: '0'
+    OPENAI_API_KEY: EMPTY
+    KV_OFFLOADING: none
diff --git a/benchmarks/multi_node/srt-slurm-recipes/trtllm/glm5.2/gb300-fp4/agentic/dynamo-disagg-gb300-8p1d-dep16-c259-b16-mtp3.yaml b/benchmarks/multi_node/srt-slurm-recipes/trtllm/glm5.2/gb300-fp4/agentic/dynamo-disagg-gb300-8p1d-dep16-c259-b16-mtp3.yaml
new file mode 100644
index 0000000000..f291bda5c4
--- /dev/null
+++ b/benchmarks/multi_node/srt-slurm-recipes/trtllm/glm5.2/gb300-fp4/agentic/dynamo-disagg-gb300-8p1d-dep16-c259-b16-mtp3.yaml
@@ -0,0 +1,185 @@
+name: dynamo-disagg-gb300-8p1d-dep16-c259-b16-mtp3
+
+model:
+  path: nvidia/GLM-5.2-NVFP4
+  container: nvcr.io/nvidia/tensorrt-llm/release:1.3.0rc22.post1
+  precision: fp4
+
+dynamo:
+  install: true
+  wheel: 1.4.0.dev20260807
+  request_plane: tcp
+
+health_check:
+  max_attempts: 270
+  interval_seconds: 10
+
+resources:
+  gpu_type: gb300
+  gpus_per_node: 4
+  prefill_nodes: 8
+  prefill_workers: 8
+  gpus_per_prefill: 4
+  decode_nodes: 4
+  decode_workers: 1
+  gpus_per_decode: 16
+
+backend:
+  type: trtllm
+  prefill_environment: &server_environment
+    HF_HUB_OFFLINE: "1"
+    TRANSFORMERS_OFFLINE: "1"
+    TQDM_DISABLE: "1"
+    HF_HUB_DISABLE_PROGRESS_BARS: "1"
+    TLLM_LOG_LEVEL: INFO
+    TRTLLM_SERVER_DISABLE_GC: "1"
+    TRTLLM_WORKER_DISABLE_GC: "1"
+    TRTLLM_ENABLE_PDL: "1"
+    NCCL_GRAPH_MIXING_SUPPORT: "0"
+    MIMALLOC_PURGE_DELAY: "0"
+    PYTORCH_CUDA_ALLOC_CONF: expandable_segments:True
+    TRTLLM_SERVE_ENABLE_MSGSPEC: "1"
+    TLLM_ADP_ROUTER_MATCH_RATE_THRESHOLD: "0.10"
+    TRTLLM_KV_CACHE_TRANSFER_TIMEOUT_SEC: "600"
+    UCX_CUDA_IPC_ENABLE_MNNVL: "y"
+    UCX_RNDV_SCHEME: put_zcopy
+    TRTLLM_KVCACHE_SEND_BUFFER_COUNT: "1"
+    TRTLLM_KVCACHE_RECV_BUFFER_COUNT: "1"
+    TLLM_SPEC_DECODE_FORCE_NUM_ACCEPTED_TOKENS: "1.99"
+    DYN_TRTLLM_ENABLE_ATTENTION_DP: "1"
+    DYN_ENGINE_CONV_AFFINITY: "1"
+    DYN_TOKENIZER: fastokens
+    DYN_PUBLISH_KV_EVENTS: "0"
+  decode_environment: *server_environment
+  trtllm_config:
+    prefill:
+      attention_dp_config:
+        enable_kv_cache_aware_routing: false
+        kv_cache_routing_conversation_affinity: true
+        kv_cache_routing_max_sessions: 65536
+      cache_transceiver_config:
+        backend: NIXL
+        transceiver_runtime: PYTHON
+        kv_cache_bounce_size_mb: 5120
+        max_tokens_in_buffer: 1048576
+        kv_transfer_timeout_ms: 600000
+      cuda_graph_config: null
+      disable_overlap_scheduler: false
+      enable_attention_dp: true
+      enable_chunked_prefill: true
+      trust_remote_code: true
+      kv_cache_config:
+        dtype: fp8
+        enable_block_reuse: true
+        event_buffer_max_size: 0
+        free_gpu_memory_fraction: 0.75
+        host_cache_size: 137438953472
+        tokens_per_block: 64
+      max_batch_size: 256
+      max_num_tokens: 8192
+      max_seq_len: 1048576
+      moe_config:
+        backend: CUTEDSL
+      moe_expert_parallel_size: 4
+      num_postprocess_workers: 8
+      pipeline_parallel_size: 1
+      print_iter_log: true
+      return_perf_metrics: false
+      scheduler_config:
+        capacity_scheduler_policy: MAX_UTILIZATION
+      sparse_attention_config:
+        algorithm: dsa
+        enable_heuristic_topk: true
+      speculative_config:
+        decoding_type: MTP
+        max_draft_len: 3
+      tensor_parallel_size: 4
+    decode:
+      cache_transceiver_config:
+        backend: NIXL
+        transceiver_runtime: PYTHON
+        kv_cache_bounce_size_mb: 5120
+        max_tokens_in_buffer: 1048576
+        kv_transfer_timeout_ms: 600000
+      cuda_graph_config:
+        batch_sizes: [1, 2, 4, 8, 16]
+        enable_padding: true
+      enable_attention_dp: true
+      enable_lm_head_tp_in_adp: false
+      trust_remote_code: true
+      kv_cache_config:
+        dtype: fp8
+        enable_block_reuse: false
+        event_buffer_max_size: 0
+        free_gpu_memory_fraction: 0.9
+        tokens_per_block: 64
+      max_batch_size: 16
+      max_num_tokens: 128
+      max_seq_len: 1048576
+      moe_config:
+        backend: CUTEDSL
+        use_low_precision_moe_combine: true
+      moe_expert_parallel_size: 16
+      num_postprocess_workers: 4
+      pipeline_parallel_size: 1
+      print_iter_log: true
+      return_perf_metrics: false
+      sparse_attention_config:
+        algorithm: dsa
+        enable_heuristic_topk: true
+        use_cute_dsl_paged_mqa_logits: true
+        use_cute_dsl_topk: true
+      speculative_config:
+        decoding_type: MTP
+        max_draft_len: 3
+      stream_interval: 20
+      tensor_parallel_size: 16
+  publish_events_and_metrics: false
+
+frontend:
+  type: dynamo
+  enable_multiple_frontends: false
+  env:
+    ETCD_LEASE_TTL: '120'
+    DYN_ROUTER_QUEUE_THRESHOLD: None
+    DYN_ROUTER_TEMPERATURE: '0'
+    DYN_TOKENIZER_CACHE: '1'
+    DYN_TOKENIZER_CACHE_BYTES: '8000000000'
+    DYN_TOKENIZER: fastokens
+    DYN_TCP_REQUEST_TIMEOUT: '30'
+    DYN_LOG: warn
+    DYN_ROUTER_SESSION_AFFINITY_TTL_SECS: '14400'
+  args:
+    router-mode: kv
+    no-kv-events: true
+    active-decode-blocks-threshold: None
+    active-prefill-tokens-threshold: None
+    active-prefill-tokens-threshold-frac: None
+  orchestrator_placement: head
+
+benchmark:
+  type: custom
+  client_placement: last_decode
+  command: bash /infmax-workspace/benchmarks/multi_node/agentic_srt.sh
+  env:
+    INFMAX_CONTAINER_WORKSPACE: /infmax-workspace
+    RESULT_DIR: /logs/agentic
+    PORT: '8000'
+    IS_MULTINODE: 'true'
+    AIPERF_DYNAMO_SESSION_TIMEOUT_SECONDS: '14400'
+    AIPERF_DATASET_MMAP_CACHE_DIR: /aiperf_mmap_cache
+    HF_HUB_CACHE: /hf_hub_cache
+    WEKA_LOADER_OVERRIDE: semianalysis_cc_traces_weka_062126
+    AIPERF_DATASET_WEKA_LIVE_ASSISTANT_RESPONSES: "0"
+    SERVED_MODEL_NAME: GLM-5.2-NVFP4
+    MAX_MODEL_LEN: '1048576'
+    AIPERF_HTTP_X_DYNAMO_SESSION_ID_FROM_CORRELATION_ID: 'true'
+    AIPERF_USE_DYNAMO_CONV_AWARE_ROUTING: '0'
+    OPENAI_API_KEY: EMPTY
+    KV_OFFLOADING: none
+    MODEL: nvidia/GLM-5.2-NVFP4
+    MODEL_PREFIX: glm5.2
+    FRAMEWORK: dynamo-trt
+    PRECISION: fp4
+    CONC: '259'
+    DURATION: '3600'
diff --git a/configs/nvidia-master.yaml b/configs/nvidia-master.yaml
index 50ed9b4713..0b994f3c6a 100644
--- a/configs/nvidia-master.yaml
+++ b/configs/nvidia-master.yaml
@@ -8843,6 +8843,151 @@ glm5.2-fp4-gb300-dynamo-sglang-agentic-disagg:
           ep: 16
           dp-attn: true
 
+# GLM-5.2 NVFP4 TensorRT-LLM AgentX on GB300 with Dynamo aggregated serving.
+# One TP8 worker serves prefill and decode on the same eight GPUs at concurrency 1.
+glm5.2-fp4-gb300-dynamo-trt-agentic-agg-mtp:
+  image: nvcr.io/nvidia/tensorrt-llm/release:1.3.0rc22.post1
+  model: nvidia/GLM-5.2-NVFP4
+  model-prefix: glm5.2
+  runner: cluster:gb300-nv
+  precision: fp4
+  framework: dynamo-trt
+  router: { name: dynamo-router, version: "1.4.0.dev20260807" }
+  multinode: true
+  disagg: false
+  scenarios:
+    agentic-coding:
+    - search-space:
+      - spec-decoding: mtp
+        conc-list: [1]
+        kv-offloading: none
+        prefill:
+          num-worker: 1
+          tp: 8
+          ep: 1
+          dp-attn: false
+          additional-settings:
+          - "CONFIG_FILE=benchmarks/multi_node/srt-slurm-recipes/trtllm/glm5.2/gb300-fp4/agentic/dynamo-agg-gb300-tp8-c1-b2-mtp8.yaml"
+          - "SLURM_PARTITION=batch_3"
+        decode:
+          num-worker: 0
+          tp: 8
+          ep: 1
+          dp-attn: false
+
+# GLM-5.2 NVFP4 TensorRT-LLM AgentX on GB300 with Dynamo disaggregated serving.
+# Six topology variants use NIXL KV transfer and MTP3/MTP5 decoding.
+glm5.2-fp4-gb300-dynamo-trt-agentic-disagg-mtp:
+  image: nvcr.io/nvidia/tensorrt-llm/release:1.3.0rc22.post1
+  model: nvidia/GLM-5.2-NVFP4
+  model-prefix: glm5.2
+  runner: cluster:gb300-nv
+  precision: fp4
+  framework: dynamo-trt
+  router: { name: dynamo-router, version: "1.4.0.dev20260807" }
+  kv-p2p-transfer: nixl
+  multinode: true
+  disagg: true
+  scenarios:
+    agentic-coding:
+    - search-space:
+      - spec-decoding: mtp
+        conc-list: [30]
+        kv-offloading: none
+        prefill:
+          num-worker: 1
+          tp: 4
+          ep: 4
+          dp-attn: true
+          additional-settings:
+          - "CONFIG_FILE=benchmarks/multi_node/srt-slurm-recipes/trtllm/glm5.2/gb300-fp4/agentic/dynamo-disagg-gb300-1p4d-tep4-c30-b2-mtp5.yaml"
+          - "SLURM_PARTITION=batch_3"
+        decode:
+          num-worker: 4
+          tp: 4
+          ep: 4
+          dp-attn: false
+      - spec-decoding: mtp
+        conc-list: [20]
+        kv-offloading: none
+        prefill:
+          num-worker: 1
+          tp: 4
+          ep: 4
+          dp-attn: true
+          additional-settings:
+          - "CONFIG_FILE=benchmarks/multi_node/srt-slurm-recipes/trtllm/glm5.2/gb300-fp4/agentic/dynamo-disagg-gb300-1p1d-tep8-c20-b5-mtp5.yaml"
+          - "SLURM_PARTITION=batch_3"
+        decode:
+          num-worker: 1
+          tp: 8
+          ep: 8
+          dp-attn: false
+      - spec-decoding: mtp
+        conc-list: [60]
+        kv-offloading: none
+        prefill:
+          num-worker: 3
+          tp: 4
+          ep: 4
+          dp-attn: true
+          additional-settings:
+          - "CONFIG_FILE=benchmarks/multi_node/srt-slurm-recipes/trtllm/glm5.2/gb300-fp4/agentic/dynamo-disagg-gb300-3p4d-tep4-c60-b5-mtp5.yaml"
+          - "SLURM_PARTITION=batch_3"
+        decode:
+          num-worker: 4
+          tp: 4
+          ep: 4
+          dp-attn: false
+      - spec-decoding: mtp
+        conc-list: [152]
+        kv-offloading: none
+        prefill:
+          num-worker: 5
+          tp: 4
+          ep: 4
+          dp-attn: true
+          additional-settings:
+          - "CONFIG_FILE=benchmarks/multi_node/srt-slurm-recipes/trtllm/glm5.2/gb300-fp4/agentic/dynamo-disagg-gb300-5p1d-dep16-c152-b16-mtp3.yaml"
+          - "SLURM_PARTITION=batch_3"
+        decode:
+          num-worker: 1
+          tp: 16
+          ep: 16
+          dp-attn: true
+      - spec-decoding: mtp
+        conc-list: [259]
+        kv-offloading: none
+        prefill:
+          num-worker: 8
+          tp: 4
+          ep: 4
+          dp-attn: true
+          additional-settings:
+          - "CONFIG_FILE=benchmarks/multi_node/srt-slurm-recipes/trtllm/glm5.2/gb300-fp4/agentic/dynamo-disagg-gb300-8p1d-dep16-c259-b16-mtp3.yaml"
+          - "SLURM_PARTITION=batch_3"
+        decode:
+          num-worker: 1
+          tp: 16
+          ep: 16
+          dp-attn: true
+      - spec-decoding: mtp
+        conc-list: [227]
+        kv-offloading: none
+        prefill:
+          num-worker: 6
+          tp: 4
+          ep: 4
+          dp-attn: true
+          additional-settings:
+          - "CONFIG_FILE=benchmarks/multi_node/srt-slurm-recipes/trtllm/glm5.2/gb300-fp4/agentic/dynamo-disagg-gb300-6p1d-dep8-c227-b16-mtp3.yaml"
+          - "SLURM_PARTITION=batch_3"
+        decode:
+          num-worker: 1
+          tp: 8
+          ep: 8
+          dp-attn: true
+
 qwen3.5-fp8-gb200-dynamo-sglang-mtp:
   image: lmsysorg/sglang:v0.5.14-cu130@sha256:5027e95bf6ec536856b1b52a91d1f35ff5c564ab83e8a94758a169ff09bb8df3
   model: Qwen/Qwen3.5-397B-A17B-FP8
diff --git a/perf-changelog.yaml b/perf-changelog.yaml
index 93b6a7c1b3..92fe447a81 100644
--- a/perf-changelog.yaml
+++ b/perf-changelog.yaml
@@ -6251,3 +6251,30 @@
   description:
     - "Refresh with lower stream interval to collect correct client metrics"
   pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/2686
+
+- config-keys:
+    - glm5.2-fp4-gb300-dynamo-trt-agentic-agg-mtp
+    - glm5.2-fp4-gb300-dynamo-trt-agentic-disagg-mtp
+  scenario-type:
+    - agentic-coding
+  description:
+    - "Add one aggregated TP8 and six disaggregated GLM-5.2 NVFP4 GB300 AgentX recipes for Dynamo and TensorRT-LLM with MTP speculative decoding."
+    - "Use the concurrency-1 aggregated Pareto point and disable TensorRT-LLM performance-metrics responses for all seven points."
+  pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/2657
+
+- config-keys:
+    - glm5.2-fp4-gb300-dynamo-trt-agentic-disagg-mtp
+  scenario-type:
+    - agentic-coding
+  description:
+    - "Place the Dynamo frontend on the Slurm head node for eval routing across the six disaggregated recipes."
+  pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/2657
+
+- config-keys:
+    - glm5.2-fp4-gb300-dynamo-trt-agentic-agg-mtp
+    - glm5.2-fp4-gb300-dynamo-trt-agentic-disagg-mtp
+  scenario-type:
+    - agentic-coding
+  description:
+    - "Use the batch_3 Slurm partition for the GLM-5.2 GB300 Dynamo-TRT AgentX configuration."
+  pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/2657
diff --git a/runners/launch_gb300-nv.sh b/runners/launch_gb300-nv.sh
index ff6597d0bb..d5a58c5fcc 100644
--- a/runners/launch_gb300-nv.sh
+++ b/runners/launch_gb300-nv.sh
@@ -7,7 +7,8 @@ set -exo pipefail
 # shellcheck source=runners/slurm_utils.sh
 source "$(dirname "${BASH_SOURCE[0]}")/slurm_utils.sh"
 
-export SLURM_PARTITION="batch_1"
+export SLURM_PARTITION="${SLURM_PARTITION:-batch_1}"
+export SBATCH_PARTITION="$SLURM_PARTITION"
 export SLURM_ACCOUNT="benchmark"
 export ENROOT_ROOTFS_WRITABLE=1
 
@@ -65,6 +66,10 @@ elif [[ $MODEL_PREFIX == "glm5.1" && $PRECISION == "fp4" ]]; then
     # in our GLM-5.1 sglang recipes.
     export MODEL_PATH=/scratch/models/GLM-5.1-NVFP4
     export SRT_SLURM_MODEL_PREFIX="glm-5-fp4"
+elif [[ $MODEL_PREFIX == "glm5.2" && $PRECISION == "fp4" && $FRAMEWORK == "dynamo-trt" ]]; then
+    export SERVED_MODEL_NAME="GLM-5.2-NVFP4"
+    export MODEL_PATH=/scratch/models/GLM-5.2-NVFP4
+    export SRT_SLURM_MODEL_PREFIX="nvidia/GLM-5.2-NVFP4"
 elif [[ $MODEL_PREFIX == "glm5.2" && $PRECISION == "fp4" ]]; then
     export MODEL_PATH=/scratch/models/GLM-5.2-NVFP4
     export SRT_SLURM_MODEL_PREFIX="glm-5.2-fp4"
@@ -206,6 +211,18 @@ if [[ "$IS_AGENTIC" == "1" && $FRAMEWORK == "dynamo-trt" && $MODEL_PREFIX == "qw
         find "$TRTLLM_RECIPES_DIR" -name "*.yaml" \
             -exec sed -i '/TLLM_SPEC_DECODE_FORCE_NUM_ACCEPTED_TOKENS/d' {} +
     fi
+elif [[ "$IS_AGENTIC" == "1" && $FRAMEWORK == "dynamo-trt" && $MODEL_PREFIX == "glm5.2" ]]; then
+    git clone https://github.com/NVIDIA/srt-slurm.git "$SRT_REPO_DIR"
+    cd "$SRT_REPO_DIR"
+    git checkout v1.0.38
+    TRTLLM_RECIPES_DIR="benchmarks/multi_node/srt-slurm-recipes/trtllm/glm5.2"
+    mkdir -p "$TRTLLM_RECIPES_DIR"
+    cp -rT "$GITHUB_WORKSPACE/benchmarks/multi_node/srt-slurm-recipes/trtllm/glm5.2" \
+        "$TRTLLM_RECIPES_DIR"
+    if [[ "${EVAL_ONLY:-false}" == "true" ]]; then
+        find "$TRTLLM_RECIPES_DIR" -name "*.yaml" \
+            -exec sed -i '/TLLM_SPEC_DECODE_FORCE_NUM_ACCEPTED_TOKENS/d' {} +
+    fi
 elif [[ "$IS_AGENTIC" == "1" && $FRAMEWORK == "dynamo-sglang" && $MODEL_PREFIX == "qwen3.5" ]]; then
     # Qwen3.5 agentic uses NVIDIA/srt-slurm v1.0.38: the two features the
     # cquil11 fork was pinned for are merged upstream (present in v1.0.36) —
@@ -662,6 +679,19 @@ if [[ "${RUN_EVAL:-false}" == "true" || "${EVAL_ONLY:-false}" == "true" ]]; then
     else
         echo "WARNING: RUN_EVAL=true but no eval results found at $EVAL_DIR"
     fi
+
+    # srt-slurm stages eval artifacts but does not write the metadata file
+    # consumed by score validation. Reuse the canonical metadata writer so
+    # topology and recipe identity stay aligned with the workflow inputs.
+    eval_conc_value="${EVAL_CONC:-${CONC:-1}}"
+    (
+        export IS_MULTINODE=true
+        # shellcheck source=benchmarks/benchmark_lib.sh
+        source "$GITHUB_WORKSPACE/benchmarks/benchmark_lib.sh"
+        _write_lm_eval_meta_json \
+            "$GITHUB_WORKSPACE/meta_env.json" "" "$eval_conc_value"
+    )
+    echo "Wrote meta_env.json (conc=${eval_conc_value}, prefix=${MODEL_PREFIX:-unknown})"
 fi
 
 # Snapshot logs to GITHUB_WORKSPACE BEFORE cleanup, so the EXIT trap's