From e0d970e5cb76ef52ed19a5cc83e476920ee7c8a5 Mon Sep 17 00:00:00 2001
From: Ankur-singh <ankusingh@nvidia.com>
Date: Wed, 24 Jun 2026 19:03:44 -0700
Subject: [PATCH 1/3] Add MiniMax-M3 NVFP4 B300 single-node vLLM benchmark with
 EAGLE3 spec decode

New minimaxm3-fp4-b300-vllm-mtp config (fp4 vLLM aggregated on b300 with EAGLE3
speculative decoding, 3 draft tokens via Inferact/MiniMax-M3-EAGLE3). The
benchmark script overlays vllm-project/vllm PR #46380 (MiniMax-M3 modelopt NVFP4
support, commit 6c08558) before serve and routes prompts through the chat
template. Target weights are pre-staged read-only at /scratch/models/MiniMax-M3-NVFP4
(added MiniMax-M3-NVFP4 to launch_b300-nv.sh STAGED_MODELS); the EAGLE3 draft
downloads to the writable /data/models.
---
 .github/configs/nvidia-master.yaml            |  37 ++++++
 .../fixed_seq_len/minimaxm3_fp4_b300_mtp.sh   | 125 ++++++++++++++++++
 perf-changelog.yaml                           |   9 ++
 runners/launch_b300-nv.sh                     |   1 +
 4 files changed, 172 insertions(+)
 create mode 100755 benchmarks/single_node/fixed_seq_len/minimaxm3_fp4_b300_mtp.sh

diff --git a/.github/configs/nvidia-master.yaml b/.github/configs/nvidia-master.yaml
index 77833a1b3..79f46c942 100644
--- a/.github/configs/nvidia-master.yaml
+++ b/.github/configs/nvidia-master.yaml
@@ -12295,6 +12295,43 @@ minimaxm3-fp8-b300-vllm:
       - { tp: 4, ep: 4, dp-attn: true, conc-start: 64, conc-end: 128 }
       - { tp: 8, ep: 8, dp-attn: true, conc-start: 128, conc-end: 512 }
 
+# EAGLE3 speculative-decoding (spec-decoding: mtp) variant of MiniMax-M3 NVFP4
+# (nvidia/MiniMax-M3-NVFP4) B300 single-node vLLM, pairing the target with the
+# Inferact/MiniMax-M3-EAGLE3 draft head (3 speculative tokens). The benchmark
+# script overlays vllm-project/vllm PR #46380 (MiniMax-M3 modelopt NVFP4
+# support, commit 6c08558) before serve and routes prompts through the chat
+# template. Target weights are pre-staged read-only at
+# /scratch/models/MiniMax-M3-NVFP4 (added to the STAGED_MODELS allow-list in
+# launch_b300-nv.sh); the EAGLE3 draft is downloaded to the writable models dir.
+minimaxm3-fp4-b300-vllm-mtp:
+  image: vllm/vllm-openai:vllm-minimax-m3-perf-x86_64-13.0.1-7a67223
+  model: nvidia/MiniMax-M3-NVFP4
+  model-prefix: minimaxm3
+  runner: b300
+  precision: fp4
+  framework: vllm
+  multinode: false
+  scenarios:
+    fixed-seq-len:
+    - isl: 1024
+      osl: 1024
+      search-space:
+      - { tp: 8, conc-start: 1, conc-end: 64, spec-decoding: mtp }
+      - { tp: 8, ep: 8, conc-start: 1, conc-end: 256, spec-decoding: mtp }
+      - { tp: 4, conc-start: 1, conc-end: 64, spec-decoding: mtp }
+      - { tp: 4, ep: 4, conc-start: 64, conc-end: 256, spec-decoding: mtp }
+      - { tp: 4, ep: 4, dp-attn: true, conc-start: 128, conc-end: 512, spec-decoding: mtp }
+      - { tp: 8, ep: 8, dp-attn: true, conc-start: 256, conc-end: 512, spec-decoding: mtp }
+    - isl: 8192
+      osl: 1024
+      search-space:
+      - { tp: 8, conc-start: 1, conc-end: 64, spec-decoding: mtp }
+      - { tp: 8, ep: 8, conc-start: 1, conc-end: 256, spec-decoding: mtp }
+      - { tp: 4, conc-start: 1, conc-end: 64, spec-decoding: mtp }
+      - { tp: 4, ep: 4, conc-start: 64, conc-end: 256, spec-decoding: mtp }
+      - { tp: 4, ep: 4, dp-attn: true, conc-start: 64, conc-end: 128, spec-decoding: mtp }
+      - { tp: 8, ep: 8, dp-attn: true, conc-start: 128, conc-end: 256, spec-decoding: mtp }
+
 # MiniMax-M3 day-zero (https://recipes.vllm.ai/MiniMaxAI/MiniMax-M3).
 # 427B total / 26B active MoE with MSA sparse attention; MXFP8 checkpoint
 # (MiniMaxAI/MiniMax-M3-MXFP8, ~444 GB) quantized by NVIDIA — native MX tensor
diff --git a/benchmarks/single_node/fixed_seq_len/minimaxm3_fp4_b300_mtp.sh b/benchmarks/single_node/fixed_seq_len/minimaxm3_fp4_b300_mtp.sh
new file mode 100755
index 000000000..1ba12522d
--- /dev/null
+++ b/benchmarks/single_node/fixed_seq_len/minimaxm3_fp4_b300_mtp.sh
@@ -0,0 +1,125 @@
+#!/usr/bin/env bash
+
+# MiniMax-M3 NVFP4 B300 single-node vLLM recipe with EAGLE3 speculative
+# decoding — same shape as minimaxm3_fp8_b300_mtp.sh but uses the
+# nvidia/MiniMax-M3-NVFP4 checkpoint. Applies vllm-project/vllm PR #46380
+# (MiniMax-M3 modelopt NVFP4 support) from commit 6c08558 by overwriting the
+# 3 changed files in the installed vLLM package before the server starts.
+
+source "$(dirname "$0")/../../benchmark_lib.sh"
+
+check_env_vars \
+    MODEL \
+    TP \
+    EP_SIZE \
+    DP_ATTENTION \
+    CONC \
+    ISL \
+    OSL \
+    MAX_MODEL_LEN \
+    RANDOM_RANGE_RATIO \
+    RESULT_FILENAME
+
+# Apply vllm-project/vllm PR #46380 (Add MiniMax-M3 modelopt NVFP4 support, commit 6c08558).
+# This patch is required for nvidia/MiniMax-M3-NVFP4: without it vLLM does not
+# recognise the NVFP4 quant config and falls back to an unsupported path.
+VLLM_DIR=$(python3 -c "import vllm, os; print(os.path.dirname(vllm.__file__))")
+for f in \
+  model_executor/layers/fused_moe/experts/trtllm_nvfp4_moe.py \
+  model_executor/layers/quantization/modelopt.py \
+  model_executor/layers/quantization/utils/flashinfer_utils.py
+do
+  curl -fsSL "https://raw.githubusercontent.com/vllm-project/vllm/6c08558/vllm/${f}" -o "${VLLM_DIR}/${f}"
+done
+python3 -c "from vllm.model_executor.layers.fused_moe.experts.trtllm_nvfp4_moe import TrtLlmNvFp4ExpertsModular; print('[nvfp4-patch] OK')"
+
+DRAFT_MODEL="Inferact/MiniMax-M3-EAGLE3"
+
+# The target weights are launched from MODEL_PATH (the b300 launcher points it
+# at the pre-staged read-only /scratch/models/MiniMax-M3-NVFP4). The EAGLE3
+# draft is not pre-staged and must be downloaded, so it cannot live next to the
+# read-only target — fetch it into the writable models dir (/data/models)
+# instead. When MODEL_PATH is unset (stand-alone runs) fall back to the HF cache.
+if [[ -n "${MODEL_PATH:-}" ]]; then
+    if [[ ! -d "$MODEL_PATH" || -z "$(ls -A "$MODEL_PATH" 2>/dev/null)" ]]; then
+        hf download "$MODEL" --local-dir "$MODEL_PATH"
+    fi
+    DRAFT_MODEL_PATH="/data/models/${DRAFT_MODEL##*/}"
+    if [[ ! -d "$DRAFT_MODEL_PATH" || -z "$(ls -A "$DRAFT_MODEL_PATH" 2>/dev/null)" ]]; then
+        hf download "$DRAFT_MODEL" --local-dir "$DRAFT_MODEL_PATH"
+    fi
+else
+    hf download "$MODEL"
+    export MODEL_PATH="$MODEL"
+    hf download "$DRAFT_MODEL"
+    DRAFT_MODEL_PATH="$DRAFT_MODEL"
+fi
+
+if [[ -n "$SLURM_JOB_ID" ]]; then
+  echo "JOB $SLURM_JOB_ID running on $SLURMD_NODENAME"
+fi
+
+nvidia-smi
+
+SERVER_LOG=/workspace/server.log
+
+export VLLM_ENGINE_READY_TIMEOUT_S=3600
+export VLLM_FLOAT32_MATMUL_PRECISION=high
+
+if [ "${DP_ATTENTION}" = "true" ]; then
+  PARALLEL_ARGS="--tensor-parallel-size=1 --data-parallel-size=$TP --enable-expert-parallel"
+elif [ "$EP_SIZE" -gt 1 ]; then
+  PARALLEL_ARGS="--tensor-parallel-size=$TP --enable-expert-parallel"
+else
+  PARALLEL_ARGS="--tensor-parallel-size=$TP"
+fi
+
+# use 3 speculative tokens for all configs for now
+NUM_SPEC_TOKENS=3
+
+if [ "${EVAL_ONLY}" = "true" ]; then
+    setup_eval_context
+    MAX_MODEL_LEN="$EVAL_MAX_MODEL_LEN"
+fi
+start_gpu_monitor
+
+set -x
+vllm serve "$MODEL_PATH" --served-model-name "$MODEL" --host 0.0.0.0 --port $PORT \
+$PARALLEL_ARGS \
+--gpu-memory-utilization 0.90 \
+--max-model-len $MAX_MODEL_LEN \
+--block-size 128 \
+--language-model-only \
+--max-cudagraph-capture-size 2048 \
+--max-num-batched-tokens "$((ISL * 2 ))" \
+--speculative-config "{\"method\": \"eagle3\", \"model\": \"$DRAFT_MODEL_PATH\", \"num_speculative_tokens\": $NUM_SPEC_TOKENS, \"attention_backend\": \"FLASH_ATTN\"}" \
+--stream-interval 20 --no-enable-prefix-caching \
+--trust-remote-code > $SERVER_LOG 2>&1 &
+
+SERVER_PID=$!
+
+wait_for_server_ready --port "$PORT" --server-log "$SERVER_LOG" --server-pid "$SERVER_PID"
+
+pip install -q datasets pandas
+
+run_benchmark_serving \
+    --model "$MODEL" \
+    --port "$PORT" \
+    --backend vllm \
+    --input-len "$ISL" \
+    --output-len "$OSL" \
+    --random-range-ratio "$RANDOM_RANGE_RATIO" \
+    --num-prompts "$((CONC * 10))" \
+    --max-concurrency "$CONC" \
+    --result-filename "$RESULT_FILENAME" \
+    --result-dir /workspace/ \
+    --trust-remote-code \
+    --use-chat-template
+
+if [ "${RUN_EVAL}" = "true" ]; then
+    run_eval --framework lm-eval --port "$PORT"
+    append_lm_eval_summary
+fi
+
+stop_gpu_monitor
+set +x
diff --git a/perf-changelog.yaml b/perf-changelog.yaml
index 632e2f92a..a62d6323c 100644
--- a/perf-changelog.yaml
+++ b/perf-changelog.yaml
@@ -4183,3 +4183,12 @@
     - "server_atom.sh: fix _MAX_CONC assignment before cudagraph size check; gate ATOM_MOE_GU_ITLV/AITER_BF16_FP8_MOE_BOUND on DeepSeek-V4-Pro only"
     - "Search space: ISL=8192 and ISL=1024, 1P1D TP4, conc 1-512"
   pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/1927
+
+- config-keys:
+    - minimaxm3-fp4-b300-vllm-mtp
+  description:
+    - "Add MiniMax-M3 NVFP4 (nvidia/MiniMax-M3-NVFP4) B300 single-node aggregated vLLM benchmark with EAGLE3 speculative decoding (spec-decoding: mtp, 3 draft tokens via Inferact/MiniMax-M3-EAGLE3)"
+    - "Image vllm/vllm-openai:vllm-minimax-m3-perf-x86_64-13.0.1-7a67223; benchmark script overlays vllm-project/vllm PR #46380 (MiniMax-M3 modelopt NVFP4 support, commit 6c08558) before serve; prompts routed through the chat template"
+    - "Target weights pre-staged read-only at /scratch/models/MiniMax-M3-NVFP4 (added MiniMax-M3-NVFP4 to launch_b300-nv.sh STAGED_MODELS); EAGLE3 draft downloaded to the writable /data/models; --block-size 128 (MSA), --language-model-only"
+    - "Sweeps tp 4/8 with and without EP and dp-attn at 1k1k and 8k1k, conc 1-512"
+  pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/XXX
diff --git a/runners/launch_b300-nv.sh b/runners/launch_b300-nv.sh
index b4802b049..7a7008691 100644
--- a/runners/launch_b300-nv.sh
+++ b/runners/launch_b300-nv.sh
@@ -362,6 +362,7 @@ else
         MiniMax-M2.7
         MiniMax-M2.7-NVFP4
         MiniMax-M3
+        MiniMax-M3-NVFP4
         Qwen3.5-397B-A17B
         Qwen3.5-397B-A17B-FP8
         Qwen3.5-397B-A17B-NVFP4

From ec78fc5984b3cc05e14d568fb2679e6db696d698 Mon Sep 17 00:00:00 2001
From: Ankur-singh <ankusingh@nvidia.com>
Date: Wed, 24 Jun 2026 19:04:16 -0700
Subject: [PATCH 2/3] Update perf-changelog pr-link for #1929

---
 perf-changelog.yaml | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/perf-changelog.yaml b/perf-changelog.yaml
index a62d6323c..94ffa84f8 100644
--- a/perf-changelog.yaml
+++ b/perf-changelog.yaml
@@ -4191,4 +4191,4 @@
     - "Image vllm/vllm-openai:vllm-minimax-m3-perf-x86_64-13.0.1-7a67223; benchmark script overlays vllm-project/vllm PR #46380 (MiniMax-M3 modelopt NVFP4 support, commit 6c08558) before serve; prompts routed through the chat template"
     - "Target weights pre-staged read-only at /scratch/models/MiniMax-M3-NVFP4 (added MiniMax-M3-NVFP4 to launch_b300-nv.sh STAGED_MODELS); EAGLE3 draft downloaded to the writable /data/models; --block-size 128 (MSA), --language-model-only"
     - "Sweeps tp 4/8 with and without EP and dp-attn at 1k1k and 8k1k, conc 1-512"
-  pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/XXX
+  pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/1929

From 98b33d07786b80d281e978b8fa4ca7d9a563c26d Mon Sep 17 00:00:00 2001
From: Ankur-singh <ankusingh@nvidia.com>
Date: Thu, 25 Jun 2026 13:13:09 -0700
Subject: [PATCH 3/3] Drop runtime NVFP4 patch; bump perf image to ...-8b00f41

The vllm/vllm-openai:vllm-minimax-m3-perf-x86_64-13.0.1-8b00f41 image bakes in
MiniMax-M3 modelopt NVFP4 support (vllm-project/vllm PR #46380), so the EAGLE3
benchmark script no longer overwrites vLLM files at runtime.
---
 .github/configs/nvidia-master.yaml            | 10 +++++-----
 .../fixed_seq_len/minimaxm3_fp4_b300_mtp.sh   | 19 +++----------------
 perf-changelog.yaml                           |  2 +-
 3 files changed, 9 insertions(+), 22 deletions(-)

diff --git a/.github/configs/nvidia-master.yaml b/.github/configs/nvidia-master.yaml
index 2953c75ae..a36960865 100644
--- a/.github/configs/nvidia-master.yaml
+++ b/.github/configs/nvidia-master.yaml
@@ -12297,14 +12297,14 @@ minimaxm3-fp8-b300-vllm:
 
 # EAGLE3 speculative-decoding (spec-decoding: mtp) variant of MiniMax-M3 NVFP4
 # (nvidia/MiniMax-M3-NVFP4) B300 single-node vLLM, pairing the target with the
-# Inferact/MiniMax-M3-EAGLE3 draft head (3 speculative tokens). The benchmark
-# script overlays vllm-project/vllm PR #46380 (MiniMax-M3 modelopt NVFP4
-# support, commit 6c08558) before serve and routes prompts through the chat
-# template. Target weights are pre-staged read-only at
+# Inferact/MiniMax-M3-EAGLE3 draft head (3 speculative tokens). MiniMax-M3
+# modelopt NVFP4 support (vllm-project/vllm PR #46380) is baked into the perf
+# container image, so no runtime patch is needed; prompts are routed through the
+# chat template. Target weights are pre-staged read-only at
 # /scratch/models/MiniMax-M3-NVFP4 (added to the STAGED_MODELS allow-list in
 # launch_b300-nv.sh); the EAGLE3 draft is downloaded to the writable models dir.
 minimaxm3-fp4-b300-vllm-mtp:
-  image: vllm/vllm-openai:vllm-minimax-m3-perf-x86_64-13.0.1-7a67223
+  image: vllm/vllm-openai:vllm-minimax-m3-perf-x86_64-13.0.1-8b00f41
   model: nvidia/MiniMax-M3-NVFP4
   model-prefix: minimaxm3
   runner: b300
diff --git a/benchmarks/single_node/fixed_seq_len/minimaxm3_fp4_b300_mtp.sh b/benchmarks/single_node/fixed_seq_len/minimaxm3_fp4_b300_mtp.sh
index 1ba12522d..74cbcd020 100755
--- a/benchmarks/single_node/fixed_seq_len/minimaxm3_fp4_b300_mtp.sh
+++ b/benchmarks/single_node/fixed_seq_len/minimaxm3_fp4_b300_mtp.sh
@@ -2,9 +2,9 @@
 
 # MiniMax-M3 NVFP4 B300 single-node vLLM recipe with EAGLE3 speculative
 # decoding — same shape as minimaxm3_fp8_b300_mtp.sh but uses the
-# nvidia/MiniMax-M3-NVFP4 checkpoint. Applies vllm-project/vllm PR #46380
-# (MiniMax-M3 modelopt NVFP4 support) from commit 6c08558 by overwriting the
-# 3 changed files in the installed vLLM package before the server starts.
+# nvidia/MiniMax-M3-NVFP4 checkpoint. MiniMax-M3 modelopt NVFP4 support
+# (vllm-project/vllm PR #46380) is baked into the perf container image, so no
+# runtime patch is needed.
 
 source "$(dirname "$0")/../../benchmark_lib.sh"
 
@@ -20,19 +20,6 @@ check_env_vars \
     RANDOM_RANGE_RATIO \
     RESULT_FILENAME
 
-# Apply vllm-project/vllm PR #46380 (Add MiniMax-M3 modelopt NVFP4 support, commit 6c08558).
-# This patch is required for nvidia/MiniMax-M3-NVFP4: without it vLLM does not
-# recognise the NVFP4 quant config and falls back to an unsupported path.
-VLLM_DIR=$(python3 -c "import vllm, os; print(os.path.dirname(vllm.__file__))")
-for f in \
-  model_executor/layers/fused_moe/experts/trtllm_nvfp4_moe.py \
-  model_executor/layers/quantization/modelopt.py \
-  model_executor/layers/quantization/utils/flashinfer_utils.py
-do
-  curl -fsSL "https://raw.githubusercontent.com/vllm-project/vllm/6c08558/vllm/${f}" -o "${VLLM_DIR}/${f}"
-done
-python3 -c "from vllm.model_executor.layers.fused_moe.experts.trtllm_nvfp4_moe import TrtLlmNvFp4ExpertsModular; print('[nvfp4-patch] OK')"
-
 DRAFT_MODEL="Inferact/MiniMax-M3-EAGLE3"
 
 # The target weights are launched from MODEL_PATH (the b300 launcher points it
diff --git a/perf-changelog.yaml b/perf-changelog.yaml
index 0f697c72b..528b4b841 100644
--- a/perf-changelog.yaml
+++ b/perf-changelog.yaml
@@ -4195,7 +4195,7 @@
     - minimaxm3-fp4-b300-vllm-mtp
   description:
     - "Add MiniMax-M3 NVFP4 (nvidia/MiniMax-M3-NVFP4) B300 single-node aggregated vLLM benchmark with EAGLE3 speculative decoding (spec-decoding: mtp, 3 draft tokens via Inferact/MiniMax-M3-EAGLE3)"
-    - "Image vllm/vllm-openai:vllm-minimax-m3-perf-x86_64-13.0.1-7a67223; benchmark script overlays vllm-project/vllm PR #46380 (MiniMax-M3 modelopt NVFP4 support, commit 6c08558) before serve; prompts routed through the chat template"
+    - "Image vllm/vllm-openai:vllm-minimax-m3-perf-x86_64-13.0.1-8b00f41 (bakes in MiniMax-M3 modelopt NVFP4 support, vllm-project/vllm PR #46380; no runtime patch needed); prompts routed through the chat template"
     - "Target weights pre-staged read-only at /scratch/models/MiniMax-M3-NVFP4 (added MiniMax-M3-NVFP4 to launch_b300-nv.sh STAGED_MODELS); EAGLE3 draft downloaded to the writable /data/models; --block-size 128 (MSA), --language-model-only"
     - "Sweeps tp 4/8 with and without EP and dp-attn at 1k1k and 8k1k, conc 1-512"
   pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/1929