diff --git a/inferencex-e2e/benchmarks/single_node/srt-slurm-recipes/minimaxm31flash/sglang/gb300-fp4-mtp/agentic-tp2.yaml b/inferencex-e2e/benchmarks/single_node/srt-slurm-recipes/minimaxm31flash/sglang/gb300-fp4-mtp/agentic-tp2.yaml
new file mode 100644
index 0000000000..ab1788f364
--- /dev/null
+++ b/inferencex-e2e/benchmarks/single_node/srt-slurm-recipes/minimaxm31flash/sglang/gb300-fp4-mtp/agentic-tp2.yaml
@@ -0,0 +1,68 @@
+base:
+  schema: 2
+  name: minimaxm31flash-fp4-gb300-sglang-agentic-tp2
+  resources:
+    gpu_type: gb300
+    gpus_per_node: 4
+  frontend:
+    type: sglang
+    enable_multiple_frontends: false
+  observability:
+    enabled: false
+    tachometer:
+      enabled: false
+  engine: sglang
+  # Cold loads of the 250 GB checkpoint plus JIT MSA kernel builds and graph capture outlast the default.
+  health_check:
+    max_attempts: 1440
+    interval_seconds: 5
+  roles:
+    agg:
+      nodes: 1
+      workers: 1
+      gpus: 2
+      # Serve flags follow the cookbook's GB300 cells at
+      # https://docs.sglang.io/cookbook/autoregressive/MiniMax/MiniMax-M3.1.
+      args:
+        served-model-name: MiniMaxAI/MiniMax-M3.1-Flash
+        trust-remote-code: true
+        reasoning-parser: minimax-m3
+        tool-call-parser: minimax-m3
+        # Seven-token DSpark, the bundled draft's dspark_block_size. Throughput runs
+        # switch to synthetic acceptance at the golden acceptance length.
+        speculative-algorithm: DSPARK
+        speculative-dspark-block-size: 7
+        speculative-use-block-verification: true
+        cuda-graph-backend-prefill: breakable
+        tokenizer-worker-num: 8
+        watchdog-timeout: 3600
+        enable-metrics: true
+      env:
+        PYTHONNOUSERSITE: '1'
+        PYTHONUNBUFFERED: '1'
+        SGLANG_TIMEOUT_KEEP_ALIVE: '1800'
+        FMHA_SM100_ALLOW_JIT: '1'
+  benchmark:
+    env:
+      # Warmup can leave bytes unacknowledged past AIPerf's 30 s default.
+      AIPERF_HTTP_TCP_USER_TIMEOUT: '900000'
+      AIPERF_REQUIRED_SERVER_METRIC_PREFIX: 'sglang:'
+
+zip_override_tp2_hicache:
+  roles:
+    agg:
+      args:
+        tensor-parallel-size: 2
+        moe-runner-backend: flashinfer_trtllm_routed
+        chunked-prefill-size: 32768
+        # 2x CONC (AgentX subagent fan-out); the decode graph batch matches it.
+        max-running-requests: [40, 48, 56, 72]
+        cuda-graph-max-bs-decode: [40, 48, 56, 72]
+        mem-fraction-static: 0.87
+        enable-hierarchical-cache: true
+        hicache-size: 180
+        hicache-write-policy: write_back
+  benchmark:
+    env:
+      CONC: ['20', '24', '28', '36']
+      KV_OFFLOADING: dram
diff --git a/inferencex-e2e/configs/nvidia-master.yaml b/inferencex-e2e/configs/nvidia-master.yaml
index 19dfcc14a6..9e524ab5f1 100644
--- a/inferencex-e2e/configs/nvidia-master.yaml
+++ b/inferencex-e2e/configs/nvidia-master.yaml
@@ -8957,6 +8957,7 @@ minimaxm31flash-fp4-gb300-sglang-agentic-dspark:
       - { tp: 4, spec-decoding: mtp, kv-offloading: dram, kv-offload-backend: { name: hicache }, conc-list: [64], srt-recipe: agentic.yaml }
       # TP2 with L2 HiCache: 180 GB per rank at conc 32-44.
       - { tp: 2, spec-decoding: mtp, kv-offloading: dram, kv-offload-backend: { name: hicache }, conc-list: [32, 40, 44], srt-recipe: agentic.yaml }
+      - { tp: 2, spec-decoding: mtp, kv-offloading: dram, kv-offload-backend: { name: hicache }, conc-list: [20, 24, 28, 36], srt-recipe: agentic-tp2.yaml }
       # High-Throughput with L2 HiCache: 175 GB per rank at conc 64 and 128.
       - { tp: 4, ep: 4, spec-decoding: mtp, kv-offloading: dram, kv-offload-backend: { name: hicache }, conc-list: [64, 128], srt-recipe: agentic.yaml }
 
diff --git a/inferencex-e2e/perf-changelog.yaml b/inferencex-e2e/perf-changelog.yaml
index ec057d9244..8471f866d2 100644
--- a/inferencex-e2e/perf-changelog.yaml
+++ b/inferencex-e2e/perf-changelog.yaml
@@ -11244,6 +11244,14 @@
   pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/XXX
   scenario-type: [agentic-coding]
 
+- config-keys:
+    - minimaxm31flash-fp4-gb300-sglang-agentic-dspark
+  description:
+    - "Add TP2 points with DRAM KV offload (L2 HiCache) at conc 20, 24, 28 and 36 to the existing curve."
+  pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/4525
+  scenario-type: [agentic-coding]
+  append-only: true
+
 - config-keys:
     - minimaxm31flash-fp4-b200-sglang-agentic-dspark
   description: