diff --git a/benchmarks/multi_node/srt-slurm-recipes/sglang/glm5.2/b200-fp4/agentic/glm5.2-agentx-1p2d.yaml b/benchmarks/multi_node/srt-slurm-recipes/sglang/glm5.2/b200-fp4/agentic/glm5.2-agentx-1p2d.yaml
new file mode 100644
index 0000000000..dfd5c50fd0
--- /dev/null
+++ b/benchmarks/multi_node/srt-slurm-recipes/sglang/glm5.2/b200-fp4/agentic/glm5.2-agentx-1p2d.yaml
@@ -0,0 +1,174 @@
+# Agentic-coding SGLang disaggregated 1P2D recipe for GLM-5.2-NVFP4 on B200.
+# Expanded from NVIDIA/srt-slurm#314 at 2f43c324211022d367d8130e42499d579ecc90d9.
+# Golden synthetic acceptance is injected only for throughput jobs; EVAL_ONLY
+# retains real target-model verification.
+name: agentx-1p2d_dtp8-c113
+model:
+  path: glm-5.2-fp4
+  container: lmsysorg/sglang:nightly-dev-cu13-20260805-211ee642
+  precision: fp4
+identity:
+  model:
+    repo: nvidia/GLM-5.2-NVFP4
+    revision: aec724e8c7b8ee9db3b48c01c320f63f9cdaf8aa
+  container:
+    image: lmsysorg/sglang:nightly-dev-cu13-20260805-211ee642
+  frameworks:
+    dynamo: 71eb001e17fa73c742f0afe1a6ed96836cb135fd
+    sglang: nightly-dev-cu13-20260805-211ee642
+resources:
+  gpu_type: b200
+  gpus_per_node: 8
+  decode_nodes: 2
+  decode_workers: 2
+  gpus_per_decode: 8
+  gpus_per_prefill: 8
+  prefill_nodes: 1
+  prefill_workers: 1
+frontend:
+  type: dynamo
+  nginx_session_affinity: true
+  nginx_session_affinity_header: X-Dynamo-Session-ID
+  enable_multiple_frontends: true
+  env:
+    PIP_BREAK_SYSTEM_PACKAGES: '1'
+  args:
+    router-mode: kv
+    router-session-affinity-ttl-secs: '3600'
+    active-decode-blocks-threshold: None
+    active-prefill-tokens-threshold: None
+    active-prefill-tokens-threshold-frac: None
+dynamo:
+  hash: 71eb001e17fa73c742f0afe1a6ed96836cb135fd
+  install: true
+backend:
+  type: sglang
+  prefill_environment:
+    TORCH_DISTRIBUTED_DEFAULT_TIMEOUT: '1800'
+    PYTHONUNBUFFERED: '1'
+    DYN_SKIP_SGLANG_LOG_FORMATTING: '1'
+    SGLANG_DISAGGREGATION_HEARTBEAT_MAX_FAILURE: '100000'
+    SGLANG_DISAGGREGATION_BOOTSTRAP_TIMEOUT: '100000'
+    SGLANG_DISAGGREGATION_WAITING_TIMEOUT: '100000'
+    MC_TE_METRIC: 'true'
+    NCCL_CUMEM_ENABLE: '1'
+    SGLANG_USE_MESSAGE_QUEUE_BROADCASTER: '0'
+    SGLANG_DISABLE_TP_MEMORY_INBALANCE_CHECK: '1'
+    SGLANG_ENABLE_THINKING: '1'
+    SGLANG_ENABLE_UNIFIED_RADIX_TREE: '1'
+    SGLANG_HICACHE_DEBUG_LOG: '1'
+    SGLANG_HICACHE_DEBUG_SAMPLE_RATE: '16384'
+    SGLANG_REASONING_EFFORT: max
+    PIP_BREAK_SYSTEM_PACKAGES: '1'
+  decode_environment:
+    TORCH_DISTRIBUTED_DEFAULT_TIMEOUT: '1800'
+    PYTHONUNBUFFERED: '1'
+    DYN_SKIP_SGLANG_LOG_FORMATTING: '1'
+    SGLANG_DISAGGREGATION_HEARTBEAT_MAX_FAILURE: '100000'
+    SGLANG_DISAGGREGATION_BOOTSTRAP_TIMEOUT: '100000'
+    SGLANG_DISAGGREGATION_WAITING_TIMEOUT: '100000'
+    MC_TE_METRIC: 'true'
+    NCCL_CUMEM_ENABLE: '1'
+    SGLANG_USE_MESSAGE_QUEUE_BROADCASTER: '0'
+    SGLANG_DISABLE_TP_MEMORY_INBALANCE_CHECK: '1'
+    SGLANG_MOE_NVFP4_DISPATCH: '1'
+    SGLANG_NVFP4_CKPT_FP8_NEXTN_MOE: '1'
+    SGLANG_ENABLE_THINKING: '1'
+    SGLANG_CLIP_MAX_NEW_TOKENS_ESTIMATION: '8'
+    SGLANG_REASONING_EFFORT: max
+    PIP_BREAK_SYSTEM_PACKAGES: '1'
+  sglang_config:
+    prefill:
+      served-model-name: nvidia/GLM-5.2-NVFP4
+      trust-remote-code: true
+      quantization: modelopt_fp4
+      kv-cache-dtype: fp8_e4m3
+      disaggregation-mode: prefill
+      disaggregation-transfer-backend: nixl
+      tensor-parallel-size: 8
+      data-parallel-size: 8
+      expert-parallel-size: 8
+      enable-dp-attention: true
+      enable-dp-lm-head: true
+      load-balance-method: total_tokens
+      chunked-prefill-size: 65536
+      max-prefill-tokens: 8192
+      max-running-requests: 16
+      cuda-graph-max-bs: 16
+      disable-cuda-graph: true
+      nsa-prefill-backend: trtllm
+      nsa-decode-backend: trtllm
+      moe-runner-backend: flashinfer_cutlass
+      fp4-gemm-backend: flashinfer_cutlass
+      enable-flashinfer-allreduce-fusion: true
+      weight-loader-prefetch-checkpoints: true
+      model-loader-extra-config: '{"enable_multithread_load": true}'
+      mem-fraction-static: 0.8
+      context-length: 1048576
+      enable-hierarchical-cache: true
+      hicache-write-policy: write_back
+      hicache-size: 270
+      hicache-io-backend: direct
+      speculative-algorithm: EAGLE
+      speculative-num-steps: 1
+      speculative-eagle-topk: 1
+      speculative-num-draft-tokens: 2
+      enable-metrics: true
+      enable-cache-report: true
+    decode:
+      served-model-name: nvidia/GLM-5.2-NVFP4
+      trust-remote-code: true
+      quantization: modelopt_fp4
+      kv-cache-dtype: fp8_e4m3
+      disaggregation-mode: decode
+      disaggregation-transfer-backend: nixl
+      disable-radix-cache: true
+      speculative-algorithm: EAGLE
+      speculative-num-steps: 2
+      speculative-eagle-topk: 1
+      speculative-num-draft-tokens: 3
+      tensor-parallel-size: 8
+      data-parallel-size: 8
+      expert-parallel-size: 1
+      enable-dp-attention: true
+      enable-dp-lm-head: true
+      max-running-requests: 16
+      cuda-graph-max-bs: 16
+      chunked-prefill-size: 64
+      context-length: 1048576
+      nsa-prefill-backend: trtllm
+      nsa-decode-backend: trtllm
+      moe-runner-backend: flashinfer_trtllm
+      fp4-gemm-backend: flashinfer_cutlass
+      skip-tokenizer-init: true
+      stream-interval: 30
+      enable-flashinfer-allreduce-fusion: true
+      weight-loader-prefetch-checkpoints: true
+      model-loader-extra-config: '{"enable_multithread_load": true}'
+      mem-fraction-static: 0.9
+      disaggregation-decode-extra-slots: 0
+      enable-metrics: true
+      enable-cache-report: true
+health_check:
+  max_attempts: 1440
+  interval_seconds: 10
+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_HTTP_X_DYNAMO_SESSION_ID_FROM_CORRELATION_ID: 'true'
+    AIPERF_REQUIRED_SERVER_METRIC_PREFIX: 'sglang:'
+    AIPERF_DATASET_MMAP_CACHE_DIR: /aiperf_mmap_cache
+    HF_HUB_CACHE: /hf_hub_cache
+infra:
+  etcd_nats_dedicated_node: false
+  nats_max_payload_mb: 64
+sbatch_directives:
+  mem: '0'
+srun_options:
+  mem: '0'
+  container-remap-root: ''
diff --git a/benchmarks/multi_node/srt-slurm-recipes/sglang/glm5.2/b200-fp4/agentic/glm5.2-agentx-1p3d.yaml b/benchmarks/multi_node/srt-slurm-recipes/sglang/glm5.2/b200-fp4/agentic/glm5.2-agentx-1p3d.yaml
new file mode 100644
index 0000000000..be53227105
--- /dev/null
+++ b/benchmarks/multi_node/srt-slurm-recipes/sglang/glm5.2/b200-fp4/agentic/glm5.2-agentx-1p3d.yaml
@@ -0,0 +1,174 @@
+# Agentic-coding SGLang disaggregated 1P3D recipe for GLM-5.2-NVFP4 on B200.
+# Expanded from NVIDIA/srt-slurm#314 at 2f43c324211022d367d8130e42499d579ecc90d9.
+# Golden synthetic acceptance is injected only for throughput jobs; EVAL_ONLY
+# retains real target-model verification.
+name: agentx-1p3d_dtp8-c113
+model:
+  path: glm-5.2-fp4
+  container: lmsysorg/sglang:nightly-dev-cu13-20260805-211ee642
+  precision: fp4
+identity:
+  model:
+    repo: nvidia/GLM-5.2-NVFP4
+    revision: aec724e8c7b8ee9db3b48c01c320f63f9cdaf8aa
+  container:
+    image: lmsysorg/sglang:nightly-dev-cu13-20260805-211ee642
+  frameworks:
+    dynamo: 71eb001e17fa73c742f0afe1a6ed96836cb135fd
+    sglang: nightly-dev-cu13-20260805-211ee642
+resources:
+  gpu_type: b200
+  gpus_per_node: 8
+  decode_nodes: 3
+  decode_workers: 3
+  gpus_per_decode: 8
+  gpus_per_prefill: 8
+  prefill_nodes: 1
+  prefill_workers: 1
+frontend:
+  type: dynamo
+  nginx_session_affinity: true
+  nginx_session_affinity_header: X-Dynamo-Session-ID
+  enable_multiple_frontends: true
+  env:
+    PIP_BREAK_SYSTEM_PACKAGES: '1'
+  args:
+    router-mode: kv
+    router-session-affinity-ttl-secs: '3600'
+    active-decode-blocks-threshold: None
+    active-prefill-tokens-threshold: None
+    active-prefill-tokens-threshold-frac: None
+dynamo:
+  hash: 71eb001e17fa73c742f0afe1a6ed96836cb135fd
+  install: true
+backend:
+  type: sglang
+  prefill_environment:
+    TORCH_DISTRIBUTED_DEFAULT_TIMEOUT: '1800'
+    PYTHONUNBUFFERED: '1'
+    DYN_SKIP_SGLANG_LOG_FORMATTING: '1'
+    SGLANG_DISAGGREGATION_HEARTBEAT_MAX_FAILURE: '100000'
+    SGLANG_DISAGGREGATION_BOOTSTRAP_TIMEOUT: '100000'
+    SGLANG_DISAGGREGATION_WAITING_TIMEOUT: '100000'
+    MC_TE_METRIC: 'true'
+    NCCL_CUMEM_ENABLE: '1'
+    SGLANG_USE_MESSAGE_QUEUE_BROADCASTER: '0'
+    SGLANG_DISABLE_TP_MEMORY_INBALANCE_CHECK: '1'
+    SGLANG_ENABLE_THINKING: '1'
+    SGLANG_ENABLE_UNIFIED_RADIX_TREE: '1'
+    SGLANG_HICACHE_DEBUG_LOG: '1'
+    SGLANG_HICACHE_DEBUG_SAMPLE_RATE: '16384'
+    SGLANG_REASONING_EFFORT: max
+    PIP_BREAK_SYSTEM_PACKAGES: '1'
+  decode_environment:
+    TORCH_DISTRIBUTED_DEFAULT_TIMEOUT: '1800'
+    PYTHONUNBUFFERED: '1'
+    DYN_SKIP_SGLANG_LOG_FORMATTING: '1'
+    SGLANG_DISAGGREGATION_HEARTBEAT_MAX_FAILURE: '100000'
+    SGLANG_DISAGGREGATION_BOOTSTRAP_TIMEOUT: '100000'
+    SGLANG_DISAGGREGATION_WAITING_TIMEOUT: '100000'
+    MC_TE_METRIC: 'true'
+    NCCL_CUMEM_ENABLE: '1'
+    SGLANG_USE_MESSAGE_QUEUE_BROADCASTER: '0'
+    SGLANG_DISABLE_TP_MEMORY_INBALANCE_CHECK: '1'
+    SGLANG_MOE_NVFP4_DISPATCH: '1'
+    SGLANG_NVFP4_CKPT_FP8_NEXTN_MOE: '1'
+    SGLANG_ENABLE_THINKING: '1'
+    SGLANG_CLIP_MAX_NEW_TOKENS_ESTIMATION: '8'
+    SGLANG_REASONING_EFFORT: max
+    PIP_BREAK_SYSTEM_PACKAGES: '1'
+  sglang_config:
+    prefill:
+      served-model-name: nvidia/GLM-5.2-NVFP4
+      trust-remote-code: true
+      quantization: modelopt_fp4
+      kv-cache-dtype: fp8_e4m3
+      disaggregation-mode: prefill
+      disaggregation-transfer-backend: nixl
+      tensor-parallel-size: 8
+      data-parallel-size: 8
+      expert-parallel-size: 8
+      enable-dp-attention: true
+      enable-dp-lm-head: true
+      load-balance-method: total_tokens
+      chunked-prefill-size: 65536
+      max-prefill-tokens: 8192
+      max-running-requests: 16
+      cuda-graph-max-bs: 16
+      disable-cuda-graph: true
+      nsa-prefill-backend: trtllm
+      nsa-decode-backend: trtllm
+      moe-runner-backend: flashinfer_cutlass
+      fp4-gemm-backend: flashinfer_cutlass
+      enable-flashinfer-allreduce-fusion: true
+      weight-loader-prefetch-checkpoints: true
+      model-loader-extra-config: '{"enable_multithread_load": true}'
+      mem-fraction-static: 0.8
+      context-length: 1048576
+      enable-hierarchical-cache: true
+      hicache-write-policy: write_back
+      hicache-size: 270
+      hicache-io-backend: direct
+      speculative-algorithm: EAGLE
+      speculative-num-steps: 1
+      speculative-eagle-topk: 1
+      speculative-num-draft-tokens: 2
+      enable-metrics: true
+      enable-cache-report: true
+    decode:
+      served-model-name: nvidia/GLM-5.2-NVFP4
+      trust-remote-code: true
+      quantization: modelopt_fp4
+      kv-cache-dtype: fp8_e4m3
+      disaggregation-mode: decode
+      disaggregation-transfer-backend: nixl
+      disable-radix-cache: true
+      speculative-algorithm: EAGLE
+      speculative-num-steps: 2
+      speculative-eagle-topk: 1
+      speculative-num-draft-tokens: 3
+      tensor-parallel-size: 8
+      data-parallel-size: 8
+      expert-parallel-size: 1
+      enable-dp-attention: true
+      enable-dp-lm-head: true
+      max-running-requests: 16
+      cuda-graph-max-bs: 16
+      chunked-prefill-size: 64
+      context-length: 1048576
+      nsa-prefill-backend: trtllm
+      nsa-decode-backend: trtllm
+      moe-runner-backend: flashinfer_trtllm
+      fp4-gemm-backend: flashinfer_cutlass
+      skip-tokenizer-init: true
+      stream-interval: 30
+      enable-flashinfer-allreduce-fusion: true
+      weight-loader-prefetch-checkpoints: true
+      model-loader-extra-config: '{"enable_multithread_load": true}'
+      mem-fraction-static: 0.9
+      disaggregation-decode-extra-slots: 0
+      enable-metrics: true
+      enable-cache-report: true
+health_check:
+  max_attempts: 1440
+  interval_seconds: 10
+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_HTTP_X_DYNAMO_SESSION_ID_FROM_CORRELATION_ID: 'true'
+    AIPERF_REQUIRED_SERVER_METRIC_PREFIX: 'sglang:'
+    AIPERF_DATASET_MMAP_CACHE_DIR: /aiperf_mmap_cache
+    HF_HUB_CACHE: /hf_hub_cache
+infra:
+  etcd_nats_dedicated_node: false
+  nats_max_payload_mb: 64
+sbatch_directives:
+  mem: '0'
+srun_options:
+  mem: '0'
+  container-remap-root: ''
diff --git a/benchmarks/multi_node/srt-slurm-recipes/sglang/glm5.2/b200-fp4/agentic/glm5.2-agentx-agg.yaml b/benchmarks/multi_node/srt-slurm-recipes/sglang/glm5.2/b200-fp4/agentic/glm5.2-agentx-agg.yaml
new file mode 100644
index 0000000000..a4aafe4d78
--- /dev/null
+++ b/benchmarks/multi_node/srt-slurm-recipes/sglang/glm5.2/b200-fp4/agentic/glm5.2-agentx-agg.yaml
@@ -0,0 +1,117 @@
+# Agentic-coding SGLang aggregated recipe for GLM-5.2-NVFP4 on B200,
+# ported from NVIDIA/srt-slurm#314 at 2f43c324211022d367d8130e42499d579ecc90d9.
+#
+# The InferenceX matrix supplies concurrency and runs this flat recipe through
+# agentic_srt.sh. Synthetic acceptance is injected only for throughput jobs;
+# EVAL_ONLY jobs retain real target-model verification.
+name: b200-fp4-glm5.2-agentx-agg
+model:
+  path: glm-5.2-fp4
+  container: lmsysorg/sglang:nightly-dev-cu13-20260805-211ee642
+  precision: fp4
+identity:
+  model:
+    repo: nvidia/GLM-5.2-NVFP4
+    revision: aec724e8c7b8ee9db3b48c01c320f63f9cdaf8aa
+  container:
+    image: lmsysorg/sglang:nightly-dev-cu13-20260805-211ee642
+  frameworks:
+    dynamo: 71eb001e17fa73c742f0afe1a6ed96836cb135fd
+    sglang: nightly-dev-cu13-20260805-211ee642
+resources:
+  gpu_type: b200
+  gpus_per_node: 8
+  agg_nodes: 1
+  agg_workers: 1
+  gpus_per_agg: 8
+frontend:
+  type: dynamo
+  env:
+    PIP_BREAK_SYSTEM_PACKAGES: '1'
+  args:
+    router-mode: kv
+    active-decode-blocks-threshold: None
+    active-prefill-tokens-threshold: None
+    active-prefill-tokens-threshold-frac: None
+    router-session-affinity-ttl-secs: 3600
+dynamo:
+  hash: 71eb001e17fa73c742f0afe1a6ed96836cb135fd
+  install: true
+backend:
+  type: sglang
+  sglang_config:
+    aggregated:
+      served-model-name: nvidia/GLM-5.2-NVFP4
+      trust-remote-code: true
+      quantization: modelopt_fp4
+      kv-cache-dtype: fp8_e4m3
+      tensor-parallel-size: 8
+      data-parallel-size: 1
+      expert-parallel-size: 1
+      enable-dp-attention: false
+      enable-dp-lm-head: false
+      max-running-requests: 10
+      cuda-graph-max-bs: 10
+      chunked-prefill-size: 8192
+      max-prefill-tokens: 8192
+      context-length: 1048576
+      speculative-algorithm: EAGLE
+      speculative-num-steps: 3
+      speculative-eagle-topk: 1
+      speculative-num-draft-tokens: 4
+      nsa-prefill-backend: trtllm
+      nsa-decode-backend: trtllm
+      moe-runner-backend: flashinfer_trtllm
+      fp4-gemm-backend: flashinfer_trtllm
+      disable-shared-experts-fusion: true
+      enable-flashinfer-allreduce-fusion: true
+      weight-loader-prefetch-checkpoints: true
+      model-loader-extra-config: '{"enable_multithread_load": true}'
+      mem-fraction-static: 0.8
+      enable-hierarchical-cache: true
+      hicache-write-policy: write_back
+      hicache-size: 270
+      hicache-io-backend: direct
+      enable-metrics: true
+      enable-cache-report: true
+  aggregated_environment:
+    TORCH_DISTRIBUTED_DEFAULT_TIMEOUT: '1800'
+    PYTHONUNBUFFERED: '1'
+    DYN_SKIP_SGLANG_LOG_FORMATTING: '1'
+    MC_TE_METRIC: 'true'
+    NCCL_CUMEM_ENABLE: '1'
+    SGLANG_USE_MESSAGE_QUEUE_BROADCASTER: '0'
+    SGLANG_DISABLE_TP_MEMORY_INBALANCE_CHECK: '1'
+    SGLANG_ENABLE_THINKING: '1'
+    SGLANG_REASONING_EFFORT: max
+    SGLANG_ENABLE_UNIFIED_RADIX_TREE: '1'
+    SGLANG_HICACHE_DEBUG_LOG: '1'
+    SGLANG_HICACHE_DEBUG_SAMPLE_RATE: '16384'
+    SGLANG_MOE_NVFP4_DISPATCH: '1'
+    SGLANG_NVFP4_CKPT_FP8_NEXTN_MOE: '1'
+    SGLANG_CLIP_MAX_NEW_TOKENS_ESTIMATION: '8'
+    PIP_BREAK_SYSTEM_PACKAGES: '1'
+health_check:
+  max_attempts: 1440
+  interval_seconds: 10
+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: 'false'
+    TP: '8'
+    AIPERF_HTTP_X_DYNAMO_SESSION_ID_FROM_CORRELATION_ID: 'true'
+    AIPERF_REQUIRED_SERVER_METRIC_PREFIX: 'sglang:'
+    AIPERF_DATASET_MMAP_CACHE_DIR: /aiperf_mmap_cache
+    HF_HUB_CACHE: /hf_hub_cache
+infra:
+  etcd_nats_dedicated_node: false
+  nats_max_payload_mb: 64
+sbatch_directives:
+  mem: '0'
+srun_options:
+  mem: '0'
+  container-remap-root: ''
diff --git a/configs/nvidia-master.yaml b/configs/nvidia-master.yaml
index 8719e6ff83..de2de91340 100644
--- a/configs/nvidia-master.yaml
+++ b/configs/nvidia-master.yaml
@@ -8443,6 +8443,93 @@ glm5.2-fp4-b200-sglang-agentic-mtp:
       search-space:
       - { tp: 8, spec-decoding: mtp, kv-offloading: dram, kv-offload-backend: { name: hicache }, conc-list: [1, 4, 8, 12, 16] }
 
+glm5.2-fp4-b200-dynamo-sglang-agentic-agg:
+  image: lmsysorg/sglang:nightly-dev-cu13-20260805-211ee642
+  model: nvidia/GLM-5.2-NVFP4
+  model-prefix: glm5.2
+  runner: cluster:b200-nscale
+  precision: fp4
+  framework: dynamo-sglang
+  router: { name: dynamo-router, version: "71eb001e17fa73c742f0afe1a6ed96836cb135fd" }
+  multinode: true
+  disagg: false
+  scenarios:
+    agentic-coding:
+    - dram-utilization: 0.80
+      search-space:
+      - spec-decoding: mtp
+        conc-list: [2, 4, 8, 12]
+        kv-offloading: dram
+        kv-offload-backend: { name: hicache }
+        prefill:
+          num-worker: 1
+          tp: 8
+          ep: 1
+          dp-attn: false
+          additional-settings:
+          - "SYNTHETIC_ACCEPTANCE=true"
+          - "SYNTHETIC_ACCEPTANCE_LENGTH=2.99"
+          - "CONFIG_FILE=recipes/sglang/glm5.2/b200-fp4/agentic/glm5.2-agentx-agg.yaml"
+        # Aggregated serving runs prefill and decode on the same TP8 worker.
+        # Keep decode at zero so matrix metadata does not double-count GPUs.
+        decode:
+          num-worker: 0
+          tp: 8
+          ep: 1
+          dp-attn: false
+
+glm5.2-fp4-b200-dynamo-sglang-agentic-disagg:
+  image: lmsysorg/sglang:nightly-dev-cu13-20260805-211ee642
+  model: nvidia/GLM-5.2-NVFP4
+  model-prefix: glm5.2
+  runner: cluster:b200-nscale
+  precision: fp4
+  framework: dynamo-sglang
+  router: { name: dynamo-router, version: "71eb001e17fa73c742f0afe1a6ed96836cb135fd" }
+  kv-p2p-transfer: nixl
+  multinode: true
+  disagg: true
+  scenarios:
+    agentic-coding:
+    - dram-utilization: 0.80
+      search-space:
+      - spec-decoding: mtp
+        conc-list: [113]
+        kv-offloading: dram
+        kv-offload-backend: { name: hicache }
+        prefill:
+          num-worker: 1
+          tp: 8
+          ep: 8
+          dp-attn: true
+          additional-settings:
+          - "SYNTHETIC_ACCEPTANCE=true"
+          - "SYNTHETIC_ACCEPTANCE_LENGTH=2.5"
+          - "CONFIG_FILE=recipes/sglang/glm5.2/b200-fp4/agentic/glm5.2-agentx-1p2d.yaml"
+        decode:
+          num-worker: 2
+          tp: 8
+          ep: 1
+          dp-attn: true
+      - spec-decoding: mtp
+        conc-list: [113]
+        kv-offloading: dram
+        kv-offload-backend: { name: hicache }
+        prefill:
+          num-worker: 1
+          tp: 8
+          ep: 8
+          dp-attn: true
+          additional-settings:
+          - "SYNTHETIC_ACCEPTANCE=true"
+          - "SYNTHETIC_ACCEPTANCE_LENGTH=2.5"
+          - "CONFIG_FILE=recipes/sglang/glm5.2/b200-fp4/agentic/glm5.2-agentx-1p3d.yaml"
+        decode:
+          num-worker: 3
+          tp: 8
+          ep: 1
+          dp-attn: true
+
 glm5.2-fp4-gb200-dynamo-sglang-agentic-agg:
   image: lmsysorg/sglang:v0.5.17-cu130
   model: nvidia/GLM-5.2-NVFP4
diff --git a/configs/runners.yaml b/configs/runners.yaml
index 6b6fdd0c71..91184b9def 100644
--- a/configs/runners.yaml
+++ b/configs/runners.yaml
@@ -253,6 +253,16 @@ labels:
     - b200-dgxc_07
     - b200-dgxc_08
     - b200-dgxc_09
+  cluster:b200-nscale:
+    - b200-nscale-slurm_0
+    - b200-nscale-slurm_1
+    - b200-nscale-slurm_2
+    - b200-nscale-slurm_3
+    - b200-nscale-slurm_4
+    - b200-nscale-slurm_5
+    - b200-nscale-slurm_6
+    - b200-nscale-slurm_7
+    - b200-nscale-slurm_8
   cluster:b300-cw:
     - b300-cw_0
     - b300-cw_01
diff --git a/perf-changelog.yaml b/perf-changelog.yaml
index 55fc4f8671..fafd07a17c 100644
--- a/perf-changelog.yaml
+++ b/perf-changelog.yaml
@@ -6225,3 +6225,15 @@
     - "Add a TP2/EP2 HiCache branch and update the resident recipes with extra_buffer_lazy, ReplaySSM speculative verification, and checkpoint page-cache release."
     - "Select TP4 C1/C2/C4/C8/C12/C20/C24, TP2/EP2 resident C4/C8/C12/C20, and TP2/EP2 HiCache C22/C24/C28/C32/C40 with golden synthetic acceptance length 3.39."
   pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/2660
+
+- config-keys:
+    - glm5.2-fp4-b200-dynamo-sglang-agentic-agg
+    - glm5.2-fp4-b200-dynamo-sglang-agentic-disagg
+  scenario-type:
+    - agentic-coding
+  description:
+    - "Add GLM-5.2 NVFP4 B200 Dynamo-SGLang AgentX aggregate and disaggregated coverage."
+    - "Cover aggregate TP8 at concurrency 2, 4, 8, and 12, plus disaggregated 1P2D and 1P3D DEP8 at concurrency 113."
+    - "Use EAGLE MTP, golden synthetic acceptance for throughput only, HiCache DRAM offload, NIXL KV transfer for disaggregated serving, and KV-aware session affinity."
+    - "Image: lmsysorg/sglang:nightly-dev-cu13-20260805-211ee642; recipe ported from NVIDIA/srt-slurm v1.0.53 (217f94387abeddfed7149a71955dc523e07cd765)."
+  pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/2673
diff --git a/runners/launch_b200-nscale-slurm.sh b/runners/launch_b200-nscale-slurm.sh
index 4bf9531d0f..db729b6cc3 100755
--- a/runners/launch_b200-nscale-slurm.sh
+++ b/runners/launch_b200-nscale-slurm.sh
@@ -9,8 +9,9 @@
 # dgxc-only change silently alter nscale runs.
 #
 # Scope: multi-node Dynamo-vLLM DeepSeek-V4-Pro and Kimi K2.6 FP4 runs, plus
-# DeepSeek-V4-Pro FP4 Dynamo-SGLang MTP, on the b200-nscale/b200-new runner
-# labels. Anything else exits non-zero.
+# DeepSeek-V4-Pro FP4 Dynamo-SGLang MTP and GLM-5.2-NVFP4 Dynamo-SGLang
+# AgentX MTP, on the b200-nscale/b200-new runner labels. Anything else exits
+# non-zero.
 
 SLURM_PARTITION="batch_1"
 SLURM_ACCOUNT="benchmark"
@@ -46,6 +47,9 @@ elif [[ $MODEL_PREFIX == "kimik2.6" && $PRECISION == "fp4" ]]; then
 elif [[ $MODEL_PREFIX == "kimik3" && $PRECISION == "fp4" ]]; then
     export MODEL_PATH="${MODEL_PATH:-$NSCALE_MODEL_ROOT/Kimi-K3}"
     export SRT_SLURM_MODEL_PREFIX="kimik3"
+elif [[ $MODEL_PREFIX == "glm5.2" && $PRECISION == "fp4" ]]; then
+    export MODEL_PATH="${MODEL_PATH:-$NSCALE_MODEL_ROOT/GLM-5.2-NVFP4}"
+    export SRT_SLURM_MODEL_PREFIX="glm-5.2-fp4"
 else
     echo "Unsupported model prefix/precision for b200-nscale: $MODEL_PREFIX/$PRECISION" >&2
     echo "Models staged under $NSCALE_MODEL_ROOT:" >&2
@@ -54,7 +58,8 @@ else
 fi
 
 if [[ $FRAMEWORK != "dynamo-vllm" ]] &&
-   [[ $MODEL_PREFIX != "dsv4" || $PRECISION != "fp4" || $FRAMEWORK != "dynamo-sglang" || $SPEC_DECODING != "mtp" ]]; then
+   [[ $MODEL_PREFIX != "dsv4" || $PRECISION != "fp4" || $FRAMEWORK != "dynamo-sglang" || $SPEC_DECODING != "mtp" ]] &&
+   [[ $MODEL_PREFIX != "glm5.2" || $PRECISION != "fp4" || $FRAMEWORK != "dynamo-sglang" || $SPEC_DECODING != "mtp" ]]; then
     echo "Unsupported framework/configuration for b200-nscale: $MODEL_PREFIX/$PRECISION/$FRAMEWORK/$SPEC_DECODING" >&2
     exit 1
 fi
@@ -80,6 +85,16 @@ elif [[ $MODEL_PREFIX == "dsv4" && $FRAMEWORK == "dynamo-sglang" ]]; then
     git checkout 04e87fcc505d6d851451781a5499ca19a02ec2b4 || exit 1
     mkdir -p recipes/sglang/deepseek-v4
     cp -rT "$GITHUB_WORKSPACE/benchmarks/multi_node/srt-slurm-recipes/sglang/deepseek-v4" recipes/sglang/deepseek-v4
+elif [[ $MODEL_PREFIX == "glm5.2" && $FRAMEWORK == "dynamo-sglang" ]]; then
+    # NVIDIA/srt-slurm v1.0.53 is the released schema validated against the
+    # GLM-5.2 B200 AgentX aggregate and disaggregated recipes.
+    git clone --branch v1.0.53 --single-branch https://github.com/NVIDIA/srt-slurm.git "$SRT_REPO_DIR" || exit 1
+    cd "$SRT_REPO_DIR" || exit 1
+    test "$(git rev-parse HEAD)" = "217f94387abeddfed7149a71955dc523e07cd765" || exit 1
+    mkdir -p recipes/sglang/glm5.2/b200-fp4/agentic
+    cp -rT \
+        "$GITHUB_WORKSPACE/benchmarks/multi_node/srt-slurm-recipes/sglang/glm5.2/b200-fp4/agentic" \
+        recipes/sglang/glm5.2/b200-fp4/agentic
 elif [[ $MODEL_PREFIX == "dsv4" ]]; then
     git clone https://github.com/NVIDIA/srt-slurm.git "$SRT_REPO_DIR" || exit 1
     cd "$SRT_REPO_DIR" || exit 1
@@ -235,7 +250,8 @@ SRTCTL_PREFLIGHT_ARGS=()
 # These weights are staged on the Slurm compute nodes, not the login node.
 if [[ $MODEL_PREFIX == "kimik2.6" ]] ||
    [[ $MODEL_PREFIX == "kimik3" ]] ||
-   [[ $MODEL_PREFIX == "dsv4" && $FRAMEWORK == "dynamo-sglang" ]]; then
+   [[ $MODEL_PREFIX == "dsv4" && $FRAMEWORK == "dynamo-sglang" ]] ||
+   [[ $MODEL_PREFIX == "glm5.2" ]]; then
     SRTCTL_PREFLIGHT_ARGS+=(--no-preflight)
 fi