Qwen3.5 FP4 GB300 disaggregated Dynamo-SGLang: update SGLang to v0.5.14-cu130 and NVFP4-V2 / Qwen3.5 FP4 GB300 分离式 Dynamo-SGLang：将 SGLang 更新至 v0.5.14-cu130 并切换到 NVFP4-V2

benchmarks/multi_node/srt-slurm-recipes/sglang/qwen3.5/gb300-fp4/8k1k/disagg/stp/8k1k_stp_lowlat_0.yaml

@@ -1,4 +1,4 @@
-# Qwen3.5-397B-A17B-NVFP4 Disaggregated 1P1D: TP4 Prefill + TP4 Decode
+# Qwen3.5-397B-A17B-NVFP4-V2 Disaggregated 1P1D: TP4 Prefill + TP4 Decode
 # Pure tensor parallel, no expert parallel (STP)
 # 8k1k sa-bench concurrency sweep on GB300
 #
@@ -83,7 +83,7 @@ backend:
 
   sglang_config:
     prefill:
-      served-model-name: "nvidia/Qwen3.5-397B-A17B-NVFP4"
+      served-model-name: "nvidia/Qwen3.5-397B-A17B-NVFP4-V2"
       model-path: "/model/"
       trust-remote-code: true
 
@@ -125,7 +125,7 @@ backend:
       log-level: "info"
 
     decode:
-      served-model-name: "nvidia/Qwen3.5-397B-A17B-NVFP4"
+      served-model-name: "nvidia/Qwen3.5-397B-A17B-NVFP4-V2"
       model-path: "/model/"
       trust-remote-code: true
 

benchmarks/multi_node/srt-slurm-recipes/sglang/qwen3.5/gb300-fp4/8k1k/disagg/stp/8k1k_stp_maxtpt_0.yaml

@@ -1,4 +1,4 @@
-# Qwen3.5-397B-A17B-NVFP4 Disaggregated 5P1D wide-EP
+# Qwen3.5-397B-A17B-NVFP4-V2 Disaggregated 5P1D wide-EP
 # Prefill: 5 workers @ TP4/DP4/EP4 with DP-attn (per-node, DEP4)
 # Decode:  1 worker  @ TP16/DP16/EP16 with DP-attn + TBO (DEP16, 4 nodes)
 # Total: 36 GB300 GPUs (5*4 + 4*4); 8k1k concurrency 1024/2048/3072.
@@ -93,7 +93,7 @@ backend:
 
   sglang_config:
     prefill:
-      served-model-name: "nvidia/Qwen3.5-397B-A17B-NVFP4"
+      served-model-name: "nvidia/Qwen3.5-397B-A17B-NVFP4-V2"
       model-path: "/model/"
       trust-remote-code: true
 
@@ -128,7 +128,7 @@ backend:
       linear-attn-decode-backend: "flashinfer"
 
     decode:
-      served-model-name: "nvidia/Qwen3.5-397B-A17B-NVFP4"
+      served-model-name: "nvidia/Qwen3.5-397B-A17B-NVFP4-V2"
       model-path: "/model/"
       trust-remote-code: true
 

benchmarks/multi_node/srt-slurm-recipes/sglang/qwen3.5/gb300-fp4/8k1k/disagg/stp/8k1k_stp_maxtpt_1.yaml

@@ -1,4 +1,4 @@
-# Qwen3.5-397B-A17B-NVFP4 Disaggregated 6P1D wide-EP
+# Qwen3.5-397B-A17B-NVFP4-V2 Disaggregated 6P1D wide-EP
 # Prefill: 6 workers @ TP4/DP4/EP4 with DP-attn (per-node, DEP4)
 # Decode:  1 worker  @ TP16/DP16/EP16 with DP-attn + TBO (DEP16, 4 nodes)
 # Total: 40 GB300 GPUs (6*4 + 4*4); 8k1k concurrency 5120.
@@ -91,7 +91,7 @@ backend:
 
   sglang_config:
     prefill:
-      served-model-name: "nvidia/Qwen3.5-397B-A17B-NVFP4"
+      served-model-name: "nvidia/Qwen3.5-397B-A17B-NVFP4-V2"
       model-path: "/model/"
       trust-remote-code: true
 
@@ -126,7 +126,7 @@ backend:
       linear-attn-decode-backend: "flashinfer"
 
     decode:
-      served-model-name: "nvidia/Qwen3.5-397B-A17B-NVFP4"
+      served-model-name: "nvidia/Qwen3.5-397B-A17B-NVFP4-V2"
       model-path: "/model/"
       trust-remote-code: true
 

benchmarks/multi_node/srt-slurm-recipes/sglang/qwen3.5/gb300-fp4/8k1k/disagg/stp/8k1k_stp_maxtpt_2.yaml

@@ -1,4 +1,4 @@
-# Qwen3.5-397B-A17B-NVFP4 Disaggregated 7P1D wide-EP
+# Qwen3.5-397B-A17B-NVFP4-V2 Disaggregated 7P1D wide-EP
 # Prefill: 7 workers @ TP4/DP4/EP4 with DP-attn (per-node, DEP4)
 # Decode:  1 worker  @ TP16/DP16/EP16 with DP-attn + TBO (DEP16, 4 nodes)
 # Total: 44 GB300 GPUs (7*4 + 4*4); 8k1k concurrency 5120.
@@ -91,7 +91,7 @@ backend:
 
   sglang_config:
     prefill:
-      served-model-name: "nvidia/Qwen3.5-397B-A17B-NVFP4"
+      served-model-name: "nvidia/Qwen3.5-397B-A17B-NVFP4-V2"
       model-path: "/model/"
       trust-remote-code: true
 
@@ -126,7 +126,7 @@ backend:
       linear-attn-decode-backend: "flashinfer"
 
     decode:
-      served-model-name: "nvidia/Qwen3.5-397B-A17B-NVFP4"
+      served-model-name: "nvidia/Qwen3.5-397B-A17B-NVFP4-V2"
       model-path: "/model/"
       trust-remote-code: true
 

configs/nvidia-master.yaml

@@ -6223,8 +6223,8 @@ qwen3.5-fp8-h100-sglang-mtp:
 
     # ---------- 1k1k high-throughput (wide-EP TP=32 decode) ----------
 qwen3.5-fp4-gb300-dynamo-sglang:
-  image: lmsysorg/sglang:nightly-dev-cu13-20260624-b2c8f7a2
-  model: nvidia/Qwen3.5-397B-A17B-NVFP4
+  image: lmsysorg/sglang:v0.5.14-cu130
+  model: nvidia/Qwen3.5-397B-A17B-NVFP4-V2
   model-prefix: qwen3.5
   runner: gb300
   precision: fp4

perf-changelog.yaml

@@ -5755,3 +5755,10 @@
     - "8k1k: 9 STP configs spanning 1P13D DEP2+13D TP4 (conc 78), 1P4D DEP2+4D TEP16 (conc 24), 1P8D DEP2+8D TEP8 (conc 48), 2P15D DEP2+15D TP4 (conc 180/216), 3P1D DEP2+1D DEP4 (conc 1536), 7P12D DEP2+12D TEP4 (conc 1152), 8P9D DEP2+9D TEP4 (conc 1728), 10P3D DEP2+3D DEP8 (conc 4608)."
     - "Runner: launch_gb300-nv.sh updated to clone NVIDIA/srt-slurm@v1.0.29 and overlay local recipes from benchmarks/multi_node/srt-slurm-recipes/trtllm/qwen3.5/gb300-fp4/disagg/ for the dynamo-trt+qwen3.5+fp4 path, and to skip srtctl preflight for this path."
   pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/2206
+
+- config-keys:
+    - qwen3.5-fp4-gb300-dynamo-sglang
+  description:
+    - "Bump SGLang image to v0.5.14-cu130 for the Qwen3.5 FP4 GB300 disaggregated dynamo-sglang config"
+    - "Switch to the nvidia/Qwen3.5-397B-A17B-NVFP4-V2 checkpoint (model pre-staged on cluster as Qwen3.5-397B-A17B-NVFP4-V2)"
+  pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/2238

runners/launch_gb300-nv.sh

@@ -83,7 +83,7 @@ elif [[ $MODEL_PREFIX == "kimik2.5" && $PRECISION == "fp4" ]]; then
 elif [[ $MODEL_PREFIX == "qwen3.5" && $PRECISION == "fp4" ]]; then
     # SRT_SLURM_MODEL_PREFIX must match the model.path alias used in our
     # Qwen3.5 sglang recipes (qwen3.5-fp4).
-    export MODEL_PATH=/scratch/models/Qwen3.5-397B-A17B-NVFP4
+    export MODEL_PATH=/scratch/models/Qwen3.5-397B-A17B-NVFP4-V2
     export SRT_SLURM_MODEL_PREFIX="qwen3.5-fp4"
 elif [[ $MODEL_PREFIX == "qwen3.5" && $PRECISION == "fp8" ]]; then
     # SRT_SLURM_MODEL_PREFIX must match the model.path alias used in our