diff --git a/benchmarks/single_node/srt-slurm-recipes/dsv41flash/vllm/b200-fp4-mtp/agentic.yaml b/benchmarks/single_node/srt-slurm-recipes/dsv41flash/vllm/b200-fp4-mtp/agentic.yaml
index 82730b356..98a4b1b01 100644
--- a/benchmarks/single_node/srt-slurm-recipes/dsv41flash/vllm/b200-fp4-mtp/agentic.yaml
+++ b/benchmarks/single_node/srt-slurm-recipes/dsv41flash/vllm/b200-fp4-mtp/agentic.yaml
@@ -6,7 +6,7 @@ base:
   name: dsv41flash-fp4-b200-vllm-agentic
   model:
     path: hf:deepseek-ai/DeepSeek-V4.1-Flash
-    container: vllm/vllm-openai:nightly-cd10ed6f9f6b37a8ace9cf380007e66fe12ec0c3
+    container: vllm/vllm-openai:nightly-ddd6fbca148a867aad1fcab7ec72f582b9977db4
     precision: fp4
   resources:
     gpu_type: b200
@@ -24,7 +24,7 @@ base:
   # The engine may take the full VLLM_ENGINE_READY_TIMEOUT_S to become ready.
   health_check:
     interval_seconds: 10
-    max_attempts: 360
+    max_attempts: 720
   roles:
     agg:
       nodes: 1
@@ -37,6 +37,10 @@ base:
         enable-auto-tool-choice: true
         reasoning-parser: deepseek_v41
         engram-config: '{"cpu_offload":true}'
+        # FlashInfer sparse attention with the merged Blackwell sparse indexer
+        # (MXFP4 indexer KV, sparse logits) at both TP sizes; fp8 KV cache.
+        attention-config: '{"backend":"FLASHINFER_MLA_SPARSE_DSV41","indexer_kv_dtype":"mxfp4","indexer_sparse_logits":true}'
+        kv-cache-dtype: fp8
         # Five-token DSpark with probabilistic drafting. Throughput runs replace
         # block rejection with the golden acceptance length.
         speculative-config: '{"method":"dspark","num_speculative_tokens":5,"draft_sample_method":"probabilistic","rejection_sample_method":"block","enable_adaptive_verification":false}'
@@ -45,7 +49,7 @@ base:
       env:
         VLLM_USE_RUST_FRONTEND: '1'
         PYTHONUNBUFFERED: '1'
-        VLLM_ENGINE_READY_TIMEOUT_S: '3600'
+        VLLM_ENGINE_READY_TIMEOUT_S: '7200'
   benchmark:
     type: custom
     command: bash /infmax-workspace/benchmarks/srt_agentic.sh
diff --git a/configs/nvidia-master.yaml b/configs/nvidia-master.yaml
index 1ecb68d8b..ab933b8b7 100644
--- a/configs/nvidia-master.yaml
+++ b/configs/nvidia-master.yaml
@@ -8346,7 +8346,7 @@ dsv41flash-fp4-b300-vllm-agentic-dspark:
       - { tp: 2, kv-offloading: none, spec-decoding: mtp, conc-list: [2, 4, 8, 16, 32, 64, 128], srt-recipe: benchmarks/single_node/srt-slurm-recipes/dsv41flash/vllm/b300-fp4-mtp/agentic.yaml }
 
 dsv41flash-fp4-b200-vllm-agentic-dspark:
-  image: vllm/vllm-openai:nightly-cd10ed6f9f6b37a8ace9cf380007e66fe12ec0c3
+  image: vllm/vllm-openai:nightly-ddd6fbca148a867aad1fcab7ec72f582b9977db4
   model: deepseek-ai/DeepSeek-V4.1-Flash
   model-prefix: dsv41flash
   runner: cluster:b200-nscale
diff --git a/perf-changelog.yaml b/perf-changelog.yaml
index 0f361fd6d..3bb8af94a 100644
--- a/perf-changelog.yaml
+++ b/perf-changelog.yaml
@@ -8962,3 +8962,9 @@
     - "旧镜像切出于 2026-09-16，新镜像领先 441 个提交，包含当前曲线所不含的三项 Kimi-K3 正确性修复：vllm-project/vllm#51483 不再将无状态的首个 chunk 当作 decode，vllm-project/vllm#57098 修复 Kimi-K3 reasoning parser，vllm-project/vllm#57430 支持 routed expert 量化。新镜像还包含本臂在并发大于 4 时均会使用的 ROCm CPU KV 卸载改动，包括 vllm-project/vllm#57160（ROCm CPU KV 卸载改用私有 pinned 张量）与 vllm-project/vllm#50045（卸载背压检测）。"
     - "This image bump does not change the DSpark draft model data type. Only the image: line changes and kimik3_fp4_mi355x_mtp.sh is unchanged; the draft loads unmodified from the published Inferact/Kimi-K3-DSpark checkpoint via --speculative-config (model=Inferact/Kimi-K3-DSpark, method=dspark). The only dtype in that speculative-config is kv_cache_dtype=fp8, which sets the draft KV-cache storage precision, not the draft weights. No flag overrides or re-quantizes the draft-model weights, so the draft dtype is preserved from its checkpoint across this re-sweep."
   pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/3419
+
+- config-keys:
+    - dsv41flash-fp4-b200-vllm-agentic-dspark
+  description:
+    - "Pin B200 DeepSeek-V4.1-Flash vLLM to nightly ddd6fbca with FlashInfer sparse attention, MXFP4 indexer KV, sparse logits, fp8 KV cache, and a 7200 s readiness timeout."
+  pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/3459