diff --git a/models/Google/gemma-4-12B-it.yaml b/models/Google/gemma-4-12B-it.yaml
index 5d5f8d31..0e925b37 100644
--- a/models/Google/gemma-4-12B-it.yaml
+++ b/models/Google/gemma-4-12B-it.yaml
@@ -16,6 +16,8 @@ meta:
     - google/gemma-4-31B-it
   hardware:
     h100: verified
+    arc_pro_b60: verified
+    arc_pro_b70: verified
 
 model:
   model_id: "google/gemma-4-12B-it"
@@ -74,7 +76,22 @@ compatible_strategies:
   - single_node_tp
   - multi_node_tp
 
-hardware_overrides: {}
+hardware_overrides:
+  xpu:
+    extra_args:
+      - "--enforce-eager"
+      - "--attention-backend"
+      - "flash_attn"
+      - "--trust-remote-code"
+      - "--gpu-memory-utilization"
+      - "0.92"
+      - "--no-enable-prefix-caching"
+      - "--max-num-batched-tokens"
+      - "8192"
+      - "--max-model-len"
+      - "5120"
+      - "--tensor-parallel-size"
+      - "2"
 
 strategy_overrides:
   single_node_tp:
@@ -114,6 +131,7 @@ guide: |
   ### Docker (recommended)
   ```bash
   docker pull vllm/vllm-openai:gemma4-unified            # NVIDIA (CUDA 13; append -cu129 for CUDA 12.9 hosts)
+  docker pull vllm/vllm-openai-xpu:latest                # Intel XPU (B60 / B70)
   ```
   TPU images are published separately by [vllm-project/tpu-inference](https://github.com/vllm-project/tpu-inference); see the Trillium / Ironwood tpu-recipes below for the pinned tag.
 
@@ -166,6 +184,35 @@ guide: |
   ```
   On CUDA 12.9 hosts, use the `vllm/vllm-openai:gemma4-unified-cu129` tag instead.
 
+  ### Docker (Intel XPU B60 / B70, TP=2, Full-Featured)
+
+  Validated on 4× Intel Arc Pro B60 / B70 with the official vLLM XPU image.
+
+  ```bash
+  docker run --device /dev/dri \
+    -v /dev/dri/by-path:/dev/dri/by-path --shm-size=16g \
+    --privileged --ipc=host -p 8000:8000 \
+    -v ~/.cache/huggingface:/root/.cache/huggingface \
+    --entrypoint bash vllm/vllm-openai-xpu:latest \
+    -c "source /opt/intel/oneapi/setvars.sh && exec vllm serve google/gemma-4-12B-it \
+    --enforce-eager \
+    --attention-backend flash_attn \
+    --host 0.0.0.0 \
+    --port 8000 \
+    --trust-remote-code \
+    --enable-auto-tool-choice \
+    --reasoning-parser gemma4 \
+    --tool-call-parser gemma4 \
+    --chat-template examples/tool_chat_template_gemma4.jinja \
+    --limit-mm-per-prompt '{"image": 4, "audio": 1}' \
+    --async-scheduling \
+    --gpu-memory-utilization 0.92 \
+    --no-enable-prefix-caching \
+    --max-num-batched-tokens 8192 \
+    --max-model-len 5120 \
+    --tensor-parallel-size 2"
+  ```
+
   ### Docker (Cloud TPU — Trillium / Ironwood)
   TPU uses the separate `vllm/vllm-tpu` image (no pip wheel). Pull the tag specified by the upstream [Trillium](https://github.com/AI-Hypercomputer/tpu-recipes/tree/main/inference/trillium/vLLM/Gemma4) or [Ironwood](https://github.com/AI-Hypercomputer/tpu-recipes/blob/main/inference/ironwood/vLLM/Gemma4/) recipe, then run:
   ```bash
diff --git a/models/Google/gemma-4-26B-A4B-it.yaml b/models/Google/gemma-4-26B-A4B-it.yaml
index 73107c98..07e44745 100644
--- a/models/Google/gemma-4-26B-A4B-it.yaml
+++ b/models/Google/gemma-4-26B-A4B-it.yaml
@@ -37,6 +37,8 @@ meta:
     dgx_station_gb300: verified
     rtx_pro_6000: verified
     rtx_5090: verified
+    arc_pro_b60: verified
+    arc_pro_b70: verified
 
 model:
   model_id: "google/gemma-4-26B-A4B-it"
@@ -237,6 +239,21 @@ hardware_overrides:
     extra_env:
       VLLM_CPU_KVCACHE_SPACE: "40"
       VLLM_CPU_ATTN_SPLIT_KV: "0"
+  xpu:
+    extra_args:
+      - "--enforce-eager"
+      - "--attention-backend"
+      - "flash_attn"
+      - "--trust-remote-code"
+      - "--gpu-memory-utilization"
+      - "0.92"
+      - "--no-enable-prefix-caching"
+      - "--max-num-batched-tokens"
+      - "8192"
+      - "--max-model-len"
+      - "5120"
+      - "--tensor-parallel-size"
+      - "4"
 
 strategy_overrides:
   single_node_tp:
@@ -287,6 +304,7 @@ guide: |
   docker pull vllm/vllm-openai:gemma4-0505-cu130  # NVIDIA Blackwell (B200/B300, CUDA 13.0)
   docker pull vllm/vllm-openai-rocm:latest   # AMD
   docker pull vllm/vllm-openai-cpu:latest-x86_64 # For Intel Xeon 6
+  docker pull vllm/vllm-openai-xpu:latest # Intel XPU (B60 / B70)
   ```
   TPU images are published separately by [vllm-project/tpu-inference](https://github.com/vllm-project/tpu-inference); see the Trillium / Ironwood tpu-recipes below for the pinned tag.
 
@@ -375,6 +393,35 @@ guide: |
 
   For additional Intel Xeon 6 deployment details, see the Intel Software Catalog entries for [Gemma 4 26B-A4B IT](https://aiswcatalog.intel.com/models/google-gemma-4-26b-a4b-it).
 
+  ### Docker (Intel XPU B60 / B70, TP=4, Full-Featured)
+
+  Validated on 4× Intel Arc Pro B60 / B70 with the official vLLM XPU image.
+
+  ```bash
+  docker run --device /dev/dri \
+    -v /dev/dri/by-path:/dev/dri/by-path --shm-size=16g \
+    --privileged --ipc=host -p 8000:8000 \
+    -v ~/.cache/huggingface:/root/.cache/huggingface \
+    --entrypoint bash vllm/vllm-openai-xpu:latest \
+    -c "source /opt/intel/oneapi/setvars.sh && exec vllm serve google/gemma-4-26B-A4B-it \
+    --enforce-eager \
+    --attention-backend flash_attn \
+    --host 0.0.0.0 \
+    --port 8000 \
+    --trust-remote-code \
+    --enable-auto-tool-choice \
+    --reasoning-parser gemma4 \
+    --tool-call-parser gemma4 \
+    --chat-template examples/tool_chat_template_gemma4.jinja \
+    --limit-mm-per-prompt.image 4 \
+    --async-scheduling \
+    --gpu-memory-utilization 0.92 \
+    --no-enable-prefix-caching \
+    --max-num-batched-tokens 8192 \
+    --max-model-len 5120 \
+    --tensor-parallel-size 4"
+  ```
+
   ## Client Usage
 
   ### Text Generation
diff --git a/models/Google/gemma-4-31B-it.yaml b/models/Google/gemma-4-31B-it.yaml
index cf1e660b..66dd55a5 100644
--- a/models/Google/gemma-4-31B-it.yaml
+++ b/models/Google/gemma-4-31B-it.yaml
@@ -18,6 +18,8 @@ meta:
     mi355x: verified
     trillium: verified
     ironwood: verified
+    arc_pro_b60: verified
+    arc_pro_b70: verified
 
 model:
   model_id: "google/gemma-4-31B-it"
@@ -93,7 +95,22 @@ compatible_strategies:
   - single_node_tp
   - multi_node_tp
 
-hardware_overrides: {}
+hardware_overrides:
+  xpu:
+    extra_args:
+      - "--enforce-eager"
+      - "--attention-backend"
+      - "flash_attn"
+      - "--trust-remote-code"
+      - "--gpu-memory-utilization"
+      - "0.92"
+      - "--no-enable-prefix-caching"
+      - "--max-num-batched-tokens"
+      - "8192"
+      - "--max-model-len"
+      - "5120"
+      - "--tensor-parallel-size"
+      - "4"
 
 strategy_overrides:
   single_node_tp:
@@ -150,6 +167,7 @@ guide: |
   docker pull vllm/vllm-openai:gemma4-0505-cu129  # NVIDIA Hopper (H100/H200, CUDA 12.9)
   docker pull vllm/vllm-openai:gemma4-0505-cu130  # NVIDIA Blackwell (B200/B300, CUDA 13.0)
   docker pull vllm/vllm-openai-rocm:latest   # AMD
+  docker pull vllm/vllm-openai-xpu:latest # Intel XPU (B60 / B70)
   ```
   TPU images are published separately by [vllm-project/tpu-inference](https://github.com/vllm-project/tpu-inference); see the Trillium / Ironwood tpu-recipes below for the pinned tag.
 
@@ -220,6 +238,35 @@ guide: |
       --host 0.0.0.0 --port 8000
   ```
 
+  ### Docker (Intel XPU B60 / B70, TP=4, Full-Featured)
+
+  Validated on 4× Intel Arc Pro B60 / B70 with the official vLLM XPU image.
+
+  ```bash
+  docker run --device /dev/dri \
+    -v /dev/dri/by-path:/dev/dri/by-path --shm-size=16g \
+    --privileged --ipc=host -p 8000:8000 \
+    -v ~/.cache/huggingface:/root/.cache/huggingface \
+    --entrypoint bash vllm/vllm-openai-xpu:latest \
+    -c "source /opt/intel/oneapi/setvars.sh && exec vllm serve google/gemma-4-31B-it \
+    --enforce-eager \
+    --attention-backend flash_attn \
+    --host 0.0.0.0 \
+    --port 8000 \
+    --trust-remote-code \
+    --enable-auto-tool-choice \
+    --reasoning-parser gemma4 \
+    --tool-call-parser gemma4 \
+    --chat-template examples/tool_chat_template_gemma4.jinja \
+    --limit-mm-per-prompt '{"image": 4, "audio": 1}' \
+    --async-scheduling \
+    --gpu-memory-utilization 0.92 \
+    --no-enable-prefix-caching \
+    --max-num-batched-tokens 8192 \
+    --max-model-len 5120 \
+    --tensor-parallel-size 4"
+  ```
+
   ### Docker (Cloud TPU — Trillium / Ironwood)
   TPU uses the separate `vllm/vllm-tpu` image (no pip wheel). Pull the tag specified by the upstream [Trillium](https://github.com/AI-Hypercomputer/tpu-recipes/tree/main/inference/trillium/vLLM/Gemma4) or [Ironwood](https://github.com/AI-Hypercomputer/tpu-recipes/blob/main/inference/ironwood/vLLM/Gemma4/) recipe, then run:
   ```bash
diff --git a/models/Qwen/Qwen3.5-0.8B.yaml b/models/Qwen/Qwen3.5-0.8B.yaml
index 6fead71d..98903fe1 100644
--- a/models/Qwen/Qwen3.5-0.8B.yaml
+++ b/models/Qwen/Qwen3.5-0.8B.yaml
@@ -12,6 +12,9 @@ meta:
   performance_headline: "Tiny Qwen3.5 dense for edge / draft-model use"
   related_recipes:
     - "Qwen/Qwen3.5-2B"
+  hardware:
+    arc_pro_b60: verified
+    arc_pro_b70: verified
 
 model:
   model_id: "Qwen/Qwen3.5-0.8B"
@@ -53,7 +56,10 @@ variants:
 compatible_strategies:
   - single_node_tp
 
-hardware_overrides: {}
+hardware_overrides:
+  xpu:
+    extra_args:
+      - "--enforce-eager"
 
 strategy_overrides: {}
 
@@ -68,7 +74,7 @@ guide: |
   ## Prerequisites
 
   - **vLLM version:** >= 0.17.0
-  - **Hardware:** any modern GPU (>=4 GB VRAM)
+  - **Hardware:** any modern GPU (>=4 GB VRAM) or Intel Arc Pro B60/B70
 
   ### Install vLLM
 
@@ -78,6 +84,11 @@ guide: |
   uv pip install -U vllm --torch-backend=auto
   ```
 
+  ### Docker
+  ```bash
+  docker pull vllm/vllm-openai-xpu:latest  # Intel XPU (B60 / B70)
+  ```
+
   ## Launching the Server
 
   ```bash
@@ -86,6 +97,22 @@ guide: |
     --reasoning-parser qwen3
   ```
 
+  ### Docker (Intel XPU B60 / B70)
+
+  Validated on 1× Intel Arc Pro B60 / B70 (B60 24 GB, B70 32 GB per card) with the
+  official vLLM XPU image `vllm/vllm-openai-xpu:latest`.
+
+  ```bash
+  docker run --device /dev/dri \
+    -v /dev/dri/by-path:/dev/dri/by-path --shm-size=16g \
+    --privileged --ipc=host -p 8000:8000 \
+    -v ~/.cache/huggingface:/root/.cache/huggingface \
+    --entrypoint bash vllm/vllm-openai-xpu:latest \
+    -c "source /opt/intel/oneapi/setvars.sh && exec vllm serve Qwen/Qwen3.5-0.8B \
+    --reasoning-parser qwen3 \
+    --enforce-eager"
+  ```
+
   ## Client Usage
 
   ```python
diff --git a/models/Qwen/Qwen3.5-27B.yaml b/models/Qwen/Qwen3.5-27B.yaml
index e982dca0..904b6c14 100644
--- a/models/Qwen/Qwen3.5-27B.yaml
+++ b/models/Qwen/Qwen3.5-27B.yaml
@@ -9,11 +9,14 @@ meta:
   tasks:
     - multimodal
     - text
-  performance_headline: "Qwen3.5 flagship dense — single-GPU FP8 or 2x GPU BF16"
+  performance_headline: "Qwen3.5 flagship dense — single-GPU FP8 or 2x GPU or 2-4x Intel Arc Pro B60/B70 BF16"
   related_recipes:
     - "Qwen/Qwen3.5-397B-A17B"
     - "Qwen/Qwen3.5-35B-A3B"
     - "Qwen/Qwen3.5-9B"
+  hardware:
+    arc_pro_b60: verified
+    arc_pro_b70: verified
 
 model:
   model_id: "Qwen/Qwen3.5-27B"
@@ -61,7 +64,7 @@ variants:
   default:
     precision: bf16
     vram_minimum_gb: 65
-    description: "Full precision BF16 — fits on 1x H200 or 2x H100"
+    description: "Full precision BF16 — fits on 1x H200, 2x H100, 4x Intel Arc Pro B60, 2x Intel Arc Pro B70"
   fp8:
     model_id: "Qwen/Qwen3.5-27B-FP8"
     precision: fp8
@@ -77,7 +80,10 @@ compatible_strategies:
   - single_node_tp
   - multi_node_tp
 
-hardware_overrides: {}
+hardware_overrides:
+  xpu:
+    extra_args:
+      - "--enforce-eager"
 
 strategy_overrides: {}
 
@@ -93,7 +99,7 @@ guide: |
   ## Prerequisites
 
   - **vLLM version:** >= 0.17.0
-  - **Hardware (BF16):** 1x H200 or 2x H100
+  - **Hardware (BF16):** 1x H200, 2x H100, 4x Intel Arc Pro B60, 2x Intel Arc Pro B70
   - **Hardware (FP8):** single 40 GB GPU (H100/H200/L40S)
   - **Hardware (Int4):** single 24 GB GPU
 
@@ -105,6 +111,11 @@ guide: |
   uv pip install -U vllm --torch-backend=auto
   ```
 
+  ### Docker
+  ```bash
+  docker pull vllm/vllm-openai-xpu:latest  # Intel XPU (B60 / B70)
+  ```
+
   ## Launching the Server
 
   ### Single-GPU FP8
@@ -141,6 +152,24 @@ guide: |
     --enable-prefix-caching
   ```
 
+  ### Docker (Intel XPU B60 / B70)
+
+  Validated on 4× Intel Arc Pro B60 or 2x Intel Arc Pro B70 (B60 24 GB, B70 32 GB per card) with the
+  official vLLM XPU image `vllm/vllm-openai-xpu:latest`.
+
+  ```bash
+  docker run --device /dev/dri \
+    -v /dev/dri/by-path:/dev/dri/by-path --shm-size=16g \
+    --privileged --ipc=host -p 8000:8000 \
+    -v ~/.cache/huggingface:/root/.cache/huggingface \
+    --entrypoint bash vllm/vllm-openai-xpu:latest \
+    -c "source /opt/intel/oneapi/setvars.sh && exec vllm serve Qwen/Qwen3.5-27B \
+    --reasoning-parser qwen3 \
+    --tensor-parallel-size 4 \
+    --max-model-len 8192 \
+    --enforce-eager"
+  ```
+
   ## Client Usage
 
   ```python
diff --git a/models/Qwen/Qwen3.5-2B.yaml b/models/Qwen/Qwen3.5-2B.yaml
index 25a99bce..ec4ffee5 100644
--- a/models/Qwen/Qwen3.5-2B.yaml
+++ b/models/Qwen/Qwen3.5-2B.yaml
@@ -9,10 +9,13 @@ meta:
   tasks:
     - multimodal
     - text
-  performance_headline: "Edge-scale Qwen3.5 dense — fits on 8 GB GPUs"
+  performance_headline: "Edge-scale Qwen3.5 dense — fits on 8 GB GPUs or 1x Intel Arc Pro B60/B70"
   related_recipes:
     - "Qwen/Qwen3.5-4B"
     - "Qwen/Qwen3.5-0.8B"
+  hardware:
+    arc_pro_b60: verified
+    arc_pro_b70: verified
 
 model:
   model_id: "Qwen/Qwen3.5-2B"
@@ -55,12 +58,15 @@ variants:
   default:
     precision: bf16
     vram_minimum_gb: 5
-    description: "Full precision BF16 — fits on an 8 GB GPU"
+    description: "Full precision BF16 — fits on an 8 GB GPU or Intel Arc Pro B60/B70"
 
 compatible_strategies:
   - single_node_tp
 
-hardware_overrides: {}
+hardware_overrides:
+  xpu:
+    extra_args:
+      - "--enforce-eager"
 
 strategy_overrides: {}
 
@@ -75,7 +81,7 @@ guide: |
   ## Prerequisites
 
   - **vLLM version:** >= 0.17.0
-  - **Hardware:** single 8 GB GPU
+  - **Hardware:** single 8 GB GPU or Intel Arc Pro B60/B70
 
   ### Install vLLM
 
@@ -85,6 +91,11 @@ guide: |
   uv pip install -U vllm --torch-backend=auto
   ```
 
+  ### Docker
+  ```bash
+  docker pull vllm/vllm-openai-xpu:latest  # Intel XPU (B60 / B70)
+  ```
+
   ## Launching the Server
 
   ```bash
@@ -93,6 +104,22 @@ guide: |
     --reasoning-parser qwen3
   ```
 
+  ### Docker (Intel XPU B60 / B70)
+
+  Validated on 1× Intel Arc Pro B60 / B70 (B60 24 GB, B70 32 GB per card) with the
+  official vLLM XPU image `vllm/vllm-openai-xpu:latest`.
+
+  ```bash
+  docker run --device /dev/dri \
+    -v /dev/dri/by-path:/dev/dri/by-path --shm-size=16g \
+    --privileged --ipc=host -p 8000:8000 \
+    -v ~/.cache/huggingface:/root/.cache/huggingface \
+    --entrypoint bash vllm/vllm-openai-xpu:latest \
+    -c "source /opt/intel/oneapi/setvars.sh && exec vllm serve Qwen/Qwen3.5-2B \
+    --reasoning-parser qwen3 \
+    --enforce-eager"
+  ```
+
   ## Client Usage
 
   ```python
diff --git a/models/Qwen/Qwen3.5-35B-A3B.yaml b/models/Qwen/Qwen3.5-35B-A3B.yaml
index 849facd6..e6d8f026 100644
--- a/models/Qwen/Qwen3.5-35B-A3B.yaml
+++ b/models/Qwen/Qwen3.5-35B-A3B.yaml
@@ -9,12 +9,14 @@ meta:
   tasks:
     - multimodal
     - text
-  performance_headline: "Compact Qwen3.5 MoE — single-GPU FP8, 2x GPU or 2x Xeon 6 NUMA nodes BF16 serving"
+  performance_headline: "Compact Qwen3.5 MoE — single-GPU FP8, 2x GPU or 2x Xeon 6 NUMA nodes or 4x Intel Arc Pro B60/B70 BF16 serving"
   related_recipes:
     - "Qwen/Qwen3.5-397B-A17B"
     - "Qwen/Qwen3.5-122B-A10B"
   hardware:
     xeon6: verified
+    arc_pro_b60: verified
+    arc_pro_b70: verified
 
 model:
   model_id: "Qwen/Qwen3.5-35B-A3B"
@@ -62,7 +64,7 @@ variants:
   default:
     precision: bf16
     vram_minimum_gb: 84
-    description: "Full precision BF16 — fits on 1x H200, 2x H100 or 2x Xeon 6 NUMA nodes"
+    description: "Full precision BF16 — fits on 1x H200, 2x H100, 2x Xeon 6 NUMA nodes, or 4x Intel Arc Pro B60 / B70"
   fp8:
     model_id: "Qwen/Qwen3.5-35B-A3B-FP8"
     precision: fp8
@@ -82,7 +84,10 @@ compatible_strategies:
   - multi_node_dep
   - multi_node_tep
 
-hardware_overrides: {}
+hardware_overrides:
+  xpu:
+    extra_args:
+      - "--enforce-eager"
 
 strategy_overrides: {}
 
@@ -97,7 +102,7 @@ guide: |
   ## Prerequisites
 
   - **vLLM version:** >= 0.17.0
-  - **Hardware (BF16):** 1x H200, 2x H100 or 2x Xeon 6 NUMA nodes
+  - **Hardware (BF16):** 1x H200, 2x H100, 2x Xeon 6 NUMA nodes, 4x Intel Arc Pro B60 / B70
   - **Hardware (FP8):** single H100/H200
   - **Hardware (Int4):** single 24 GB GPU
 
@@ -120,6 +125,11 @@ guide: |
   uv pip install https://github.com/vllm-project/vllm/releases/download/v${VLLM_VERSION}/vllm-${VLLM_VERSION}+cpu-cp38-abi3-manylinux_2_35_x86_64.whl --torch-backend cpu
   ```
 
+  ### Docker
+  ```bash
+  docker pull vllm/vllm-openai-xpu:latest  # Intel XPU (B60 / B70)
+  ```
+
   ## Launching the Server
 
   ### Single-GPU FP8
@@ -162,6 +172,24 @@ guide: |
     --reasoning-parser qwen3
   ```
 
+  ### Docker (Intel XPU B60 / B70)
+
+  Validated on 4× Intel Arc Pro B60 / B70 (B60 24 GB, B70 32 GB per card) with the
+  official vLLM XPU image `vllm/vllm-openai-xpu:latest`.
+
+  ```bash
+  docker run --device /dev/dri \
+    -v /dev/dri/by-path:/dev/dri/by-path --shm-size=16g \
+    --privileged --ipc=host -p 8000:8000 \
+    -v ~/.cache/huggingface:/root/.cache/huggingface \
+    --entrypoint bash vllm/vllm-openai-xpu:latest \
+    -c "source /opt/intel/oneapi/setvars.sh && exec vllm serve Qwen/Qwen3.5-35B-A3B \
+    --reasoning-parser qwen3 \
+    --tensor-parallel-size 4 \
+    --max-model-len 8192 \
+    --enforce-eager"
+  ```
+
   ## Client Usage
 
   ```python
diff --git a/models/Qwen/Qwen3.5-4B.yaml b/models/Qwen/Qwen3.5-4B.yaml
index c0e86e74..30e62328 100644
--- a/models/Qwen/Qwen3.5-4B.yaml
+++ b/models/Qwen/Qwen3.5-4B.yaml
@@ -2,7 +2,7 @@ meta:
   title: "Qwen3.5-4B"
   slug: "qwen3.5-4b"
   provider: "Qwen"
-  description: "Qwen3.5 compact dense multimodal model (4B) — fits on 16 GB consumer GPUs with full 262K context or one Xeon 6 NUMA node"
+  description: "Qwen3.5 compact dense multimodal model (4B) — fits on 16 GB consumer GPUs with full 262K context or one Xeon 6 NUMA node or Intel Arc Pro B60/B70"
   date_added: 2026-04-22
   date_updated: 2026-04-22
   difficulty: beginner
@@ -15,6 +15,8 @@ meta:
     - "Qwen/Qwen3.5-2B"
   hardware:
     xeon6: verified
+    arc_pro_b60: verified
+    arc_pro_b70: verified
 
 model:
   model_id: "Qwen/Qwen3.5-4B"
@@ -57,12 +59,15 @@ variants:
   default:
     precision: bf16
     vram_minimum_gb: 10
-    description: "Full precision BF16 — fits on a single 16 GB GPU or one Xeon 6 NUMA node"
+    description: "Full precision BF16 — fits on a single 16 GB GPU or one Xeon 6 NUMA node or one Intel Arc Pro B60/B70"
 
 compatible_strategies:
   - single_node_tp
 
-hardware_overrides: {}
+hardware_overrides:
+  xpu:
+    extra_args:
+      - "--enforce-eager"
 
 strategy_overrides: {}
 
@@ -77,7 +82,7 @@ guide: |
   ## Prerequisites
 
   - **vLLM version:** >= 0.17.0
-  - **Hardware:** single 16 GB GPU (RTX 4080 / L4 / A10 / T4-24GB) or one Xeon 6 NUMA node
+  - **Hardware:** single 16 GB GPU (RTX 4080 / L4 / A10 / T4-24GB) or one Xeon 6 NUMA node or one Intel Arc Pro B60/B70
 
   ### Install vLLM
 
@@ -90,6 +95,11 @@ guide: |
   ### pip (Intel Xeon 6 CPUs)
   For Intel and AMD x86 CPUs, follow the [CPU pre-built wheels](https://docs.vllm.ai/en/latest/getting_started/installation/cpu/#pre-built-wheels) installation instructions.
 
+  ### Docker 
+  ```bash
+  docker pull vllm/vllm-openai-xpu:latest  # Intel XPU (B60 / B70)
+  ```
+
   ## Launching the Server
 
   ```bash
@@ -113,6 +123,22 @@ guide: |
       --port 8000
   ```
 
+  ### Docker (Intel XPU B60 / B70)
+
+  Validated on 1× Intel Arc Pro B60 / B70 (B60 24 GB, B70 32 GB per card) with the
+  official vLLM XPU image `vllm/vllm-openai-xpu:latest`.
+
+  ```bash
+  docker run --device /dev/dri \
+    -v /dev/dri/by-path:/dev/dri/by-path --shm-size=16g \
+    --privileged --ipc=host -p 8000:8000 \
+    -v ~/.cache/huggingface:/root/.cache/huggingface \
+    --entrypoint bash vllm/vllm-openai-xpu:latest \
+    -c "source /opt/intel/oneapi/setvars.sh && exec vllm serve Qwen/Qwen3.5-4B \
+    --reasoning-parser qwen3 \
+    --enforce-eager"
+  ```
+
 
   ### MTP speculative decoding
 
diff --git a/models/Qwen/Qwen3.5-9B.yaml b/models/Qwen/Qwen3.5-9B.yaml
index bb456348..ea9e3acf 100644
--- a/models/Qwen/Qwen3.5-9B.yaml
+++ b/models/Qwen/Qwen3.5-9B.yaml
@@ -13,6 +13,9 @@ meta:
   related_recipes:
     - "Qwen/Qwen3.5-27B"
     - "Qwen/Qwen3.5-4B"
+  hardware:
+    arc_pro_b60: verified
+    arc_pro_b70: verified
 
 model:
   model_id: "Qwen/Qwen3.5-9B"
@@ -55,13 +58,16 @@ variants:
   default:
     precision: bf16
     vram_minimum_gb: 22
-    description: "Full precision BF16 — single 24 GB GPU"
+    description: "Full precision BF16 — single 24 GB GPU or Intel Arc Pro B60/B70"
 
 compatible_strategies:
   - single_node_tp
   - multi_node_tp
 
-hardware_overrides: {}
+hardware_overrides:
+  xpu:
+    extra_args:
+      - "--enforce-eager"
 
 strategy_overrides: {}
 
@@ -76,7 +82,7 @@ guide: |
   ## Prerequisites
 
   - **vLLM version:** >= 0.17.0
-  - **Hardware:** single 24 GB GPU (RTX 4090 / L4 / A10G / H100)
+  - **Hardware:** single 24 GB GPU (RTX 4090 / L4 / A10G / H100) or Intel Arc Pro B60/B70
 
   ### Install vLLM
 
@@ -86,6 +92,11 @@ guide: |
   uv pip install -U vllm --torch-backend=auto
   ```
 
+  ### Docker
+  ```bash
+  docker pull vllm/vllm-openai-xpu:latest  # Intel XPU (B60 / B70)
+  ```
+
   ## Launching the Server
 
   ### Single-GPU BF16
@@ -104,6 +115,23 @@ guide: |
     --reasoning-parser qwen3
   ```
 
+  ### Docker (Intel XPU B60 / B70)
+
+  Validated on 1× Intel Arc Pro B60 / B70 (B60 24 GB, B70 32 GB per card) with the
+  official vLLM XPU image `vllm/vllm-openai-xpu:latest`.
+  
+  ```bash
+  docker run --device /dev/dri \
+    -v /dev/dri/by-path:/dev/dri/by-path --shm-size=16g \
+    --privileged --ipc=host -p 8000:8000 \
+    -v ~/.cache/huggingface:/root/.cache/huggingface \
+    --entrypoint bash vllm/vllm-openai-xpu:latest \
+    -c "source /opt/intel/oneapi/setvars.sh && exec vllm serve Qwen/Qwen3.5-9B \
+    --reasoning-parser qwen3 \
+    --max-model-len 8192 \
+    --enforce-eager"
+  ```
+
   ## Client Usage
 
   ```python
diff --git a/models/Qwen/Qwen3.6-27B.yaml b/models/Qwen/Qwen3.6-27B.yaml
index 6736b36f..dd3c8a4e 100644
--- a/models/Qwen/Qwen3.6-27B.yaml
+++ b/models/Qwen/Qwen3.6-27B.yaml
@@ -9,7 +9,7 @@ meta:
   tasks:
     - multimodal
     - text
-  performance_headline: "Qwen3.6 flagship dense — single-GPU FP8 or 2x GPU BF16"
+  performance_headline: "Qwen3.6 flagship dense — single-GPU FP8 or 2x GPU or 2-4x Intel Arc Pro B60/B70 BF16"
   related_recipes:
     - "Qwen/Qwen3.6-35B-A3B"
   hardware:
@@ -17,6 +17,8 @@ meta:
     rtx_pro_6000: verified
     dgx_station_gb300: verified
     rtx_5090: verified
+    arc_pro_b60: verified
+    arc_pro_b70: verified
 
 model:
   model_id: "Qwen/Qwen3.6-27B"
@@ -64,7 +66,7 @@ variants:
   default:
     precision: bf16
     vram_minimum_gb: 65
-    description: "Full precision BF16 — fits on 1x H200 or 2x H100"
+    description: "Full precision BF16 — fits on 1x H200, 2x H100, 4x Intel Arc Pro B60, 2x Intel Arc Pro B70"
     hardware_overrides:
       rtx_pro_6000:
         extra_args:
@@ -130,7 +132,10 @@ compatible_strategies:
   - single_node_tp
   - multi_node_tp
 
-hardware_overrides: {}
+hardware_overrides:
+  xpu:
+    extra_args:
+      - "--enforce-eager"
 
 strategy_overrides: {}
 
@@ -147,7 +152,7 @@ guide: |
 
   - **vLLM version:** >= 0.17.0
   - **DGX Spark NVFP4 vLLM version:** >= 0.24.0
-  - **Hardware (BF16):** 1x H200 or 2x H100
+  - **Hardware (BF16):** 1x H200, 2x H100, 4x Intel Arc Pro B60, 2x Intel Arc Pro B70
   - **Hardware (FP8):** single 40 GB GPU (H100/H200/L40S)
   - **Hardware (Int4):** single 24 GB GPU
   - **Hardware (NVFP4):** single NVIDIA Blackwell GPU (B200/B300/GB10 DGX Spark)
@@ -160,6 +165,11 @@ guide: |
   uv pip install -U vllm --torch-backend=auto
   ```
 
+  ### Docker
+  ```bash
+  docker pull vllm/vllm-openai-xpu:latest  # Intel XPU (B60 / B70)
+  ```
+
   ## Launching the Server
 
   ### Single-GPU FP8
@@ -366,6 +376,24 @@ guide: |
     --reasoning-parser qwen3
   ```
 
+  ### Docker (Intel XPU B60 / B70)
+
+  Validated on 4× Intel Arc Pro B60 or 2x Intel Arc Pro B70 (B60 24 GB, B70 32 GB per card) with the
+  official vLLM XPU image `vllm/vllm-openai-xpu:latest`.
+
+  ```bash
+  docker run --device /dev/dri \
+    -v /dev/dri/by-path:/dev/dri/by-path --shm-size=16g \
+    --privileged --ipc=host -p 8000:8000 \
+    -v ~/.cache/huggingface:/root/.cache/huggingface \
+    --entrypoint bash vllm/vllm-openai-xpu:latest \
+    -c "source /opt/intel/oneapi/setvars.sh && exec vllm serve Qwen/Qwen3.6-27B \
+    --reasoning-parser qwen3 \
+    --tensor-parallel-size 4 \
+    --max-model-len 8192 \
+    --enforce-eager"
+  ```
+
   ## Processing Ultra-Long Texts
 
   Qwen3.6-27B natively supports `262,144` tokens. For longer inputs, apply
diff --git a/models/Qwen/Qwen3.6-35B-A3B.yaml b/models/Qwen/Qwen3.6-35B-A3B.yaml
index 0308f865..c3cae464 100644
--- a/models/Qwen/Qwen3.6-35B-A3B.yaml
+++ b/models/Qwen/Qwen3.6-35B-A3B.yaml
@@ -9,7 +9,7 @@ meta:
   tasks:
     - multimodal
     - text
-  performance_headline: "Compact Qwen3.6 MoE with 3B active parameters — single-GPU FP8 or 2-4 GPU BF16 serving"
+  performance_headline: "Compact Qwen3.6 MoE with 3B active parameters — single-GPU FP8 or 2-4 GPU or 4x Intel Arc Pro B60/B70 BF16 serving"
   related_recipes:
     - "Qwen/Qwen3.5-397B-A17B"
   hardware:
@@ -22,6 +22,8 @@ meta:
     rtx_pro_6000: verified
     dgx_station_gb300: verified
     rtx_5090: verified
+    arc_pro_b60: verified
+    arc_pro_b70: verified
 
 model:
   model_id: "Qwen/Qwen3.6-35B-A3B"
@@ -69,7 +71,7 @@ variants:
   default:
     precision: bf16
     vram_minimum_gb: 84
-    description: "Full precision BF16 — fits on 1x H200 or 2x H100"
+    description: "Full precision BF16 — fits on 1x H200, 2x H100, 4x Intel Arc Pro B60 / B70"
     hardware_overrides:
       rtx_pro_6000:
         extra_args:
@@ -133,6 +135,9 @@ hardware_overrides:
     extra_args: []
     extra_env:
       VLLM_ROCM_USE_AITER: "1"
+  xpu:
+    extra_args:
+      - "--enforce-eager"
 
 strategy_overrides: {}
 
@@ -149,7 +154,7 @@ guide: |
 
   - **vLLM version:** >= 0.17.0
   - **DGX Spark NVFP4 vLLM version:** >= 0.24.0
-  - **Hardware (BF16):** 1x H200 or 2x H100
+  - **Hardware (BF16):** 1x H200, 2x H100, 4x Intel Arc Pro B60 / B70
   - **Hardware (FP8):** single H100/H200 or 1x MI300X/MI325X/MI355X
   - **Hardware (NVFP4):** NVIDIA Blackwell GPUs, including DGX Spark (GB10)
 
@@ -161,6 +166,11 @@ guide: |
   uv pip install -U vllm --torch-backend=auto
   ```
 
+  ### Docker
+  ```bash
+  docker pull vllm/vllm-openai-xpu:latest  # Intel XPU (B60 / B70)
+  ```
+
   ## Launching the Server
 
   ### Single-GPU FP8
@@ -254,6 +264,24 @@ guide: |
     --enable-auto-tool-choice
   ```
 
+  ### Docker (Intel XPU B60 / B70)
+
+  Validated on 4× Intel Arc Pro B60 / B70 (B60 24 GB, B70 32 GB per card) with the
+  official vLLM XPU image `vllm/vllm-openai-xpu:latest`.
+
+  ```bash
+  docker run --device /dev/dri \
+    -v /dev/dri/by-path:/dev/dri/by-path --shm-size=16g \
+    --privileged --ipc=host -p 8000:8000 \
+    -v ~/.cache/huggingface:/root/.cache/huggingface \
+    --entrypoint bash vllm/vllm-openai-xpu:latest \
+    -c "source /opt/intel/oneapi/setvars.sh && exec vllm serve Qwen/Qwen3.6-35B-A3B \
+    --reasoning-parser qwen3 \
+    --tensor-parallel-size 4 \
+    --max-model-len 8192 \
+    --enforce-eager"
+  ```
+
   ## Processing Ultra-Long Texts
 
   Qwen3.6-35B-A3B natively supports `262,144` tokens. For longer inputs, apply
diff --git a/models/nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16.yaml b/models/nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16.yaml
index a1e0152f..de867e94 100644
--- a/models/nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16.yaml
+++ b/models/nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16.yaml
@@ -13,6 +13,8 @@ meta:
   hardware:
     h100: verified
     h200: verified
+    arc_pro_b60: verified
+    arc_pro_b70: verified
 
 model:
   model_id: "nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16"
@@ -60,9 +62,11 @@ variants:
     extra_args:
       - "--kv-cache-dtype"
       - "fp8"
-    extra_env:
-      VLLM_USE_FLASHINFER_MOE_FP8: "1"
-      VLLM_FLASHINFER_MOE_BACKEND: "throughput"
+    hardware_overrides:
+      nvidia:
+        extra_env:
+          VLLM_USE_FLASHINFER_MOE_FP8: "1"
+          VLLM_FLASHINFER_MOE_BACKEND: "throughput"
 
 compatible_strategies:
   - single_node_tp
@@ -75,7 +79,10 @@ kv_offload_support:
   offloading_cpu: verified
   offloading_fs: verified
 
-hardware_overrides: {}
+hardware_overrides:
+  xpu:
+    extra_args:
+      - "--enforce-eager"
 
 strategy_overrides: {}
 
@@ -88,7 +95,7 @@ guide: |
 
   ## Prerequisites
 
-  - Hardware: 1x H100/H200 or comparable; DGX Spark and Jetson Thor supported
+  - Hardware: 1x H100/H200 or comparable; DGX Spark, Jetson Thor, and Intel Arc Pro B60/B70 supported
   - vLLM >= 0.11.2 (0.12.0 recommended for full support)
   - Docker with NVIDIA Container Toolkit (recommended)
 
@@ -111,6 +118,12 @@ guide: |
   docker pull ghcr.io/nvidia-ai-iot/vllm:latest-jetson-thor
   ```
 
+  Intel XPU (B60 / B70):
+
+  ```bash
+  docker pull vllm/vllm-openai-xpu:latest
+  ```
+
   ## Launch commands
 
   FP8 with FlashInfer MoE backend (Blackwell/Hopper):
@@ -142,6 +155,47 @@ guide: |
     --reasoning-parser nano_v3
   ```
 
+  ### Intel XPU (B60, B70)
+
+  Validated on 4× Intel Arc Pro B60 / B70 (B60 24 GB, B70 32 GB per card) with the
+  official vLLM XPU image `vllm/vllm-openai-xpu:latest`.
+
+  FP8:
+  ```bash
+  docker run --device /dev/dri \
+    -v /dev/dri/by-path:/dev/dri/by-path --shm-size=16g \
+    --privileged --ipc=host -p 8000:8000 \
+    -v ~/.cache/huggingface:/root/.cache/huggingface \
+    --entrypoint bash vllm/vllm-openai-xpu:latest \
+    -c "source /opt/intel/oneapi/setvars.sh && exec vllm serve nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-FP8 \
+    --trust-remote-code \
+    --async-scheduling \
+    --kv-cache-dtype fp8 \
+    --tensor-parallel-size 4 \
+    --enforce-eager \
+    --max-model-len 8192"
+  ```
+
+  BF16:
+  ```bash
+  docker run --device /dev/dri \
+    -v /dev/dri/by-path:/dev/dri/by-path --shm-size=16g \
+    --privileged --ipc=host -p 8000:8000 \
+    -v ~/.cache/huggingface:/root/.cache/huggingface \
+    --entrypoint bash vllm/vllm-openai-xpu:latest \
+    -c "source /opt/intel/oneapi/setvars.sh && exec vllm serve nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16 \
+    --trust-remote-code \
+    --async-scheduling \
+    --kv-cache-dtype auto \
+    --tensor-parallel-size 4 \
+    --enforce-eager \
+    --enable-auto-tool-choice \
+    --tool-call-parser qwen3_coder \
+    --reasoning-parser-plugin nano_v3_reasoning_parser.py \
+    --reasoning-parser nano_v3 \
+    --max-model-len 8192"
+  ```
+
   Key flags:
   - `kv-cache-dtype fp8` for FP8 variant, `auto` for BF16
   - `async-scheduling` reduces host overhead between decode steps
diff --git a/models/nvidia/NVIDIA-Nemotron-3-Nano-4B-BF16.yaml b/models/nvidia/NVIDIA-Nemotron-3-Nano-4B-BF16.yaml
index 24a535ae..7444cb21 100644
--- a/models/nvidia/NVIDIA-Nemotron-3-Nano-4B-BF16.yaml
+++ b/models/nvidia/NVIDIA-Nemotron-3-Nano-4B-BF16.yaml
@@ -15,6 +15,8 @@ meta:
     h100: verified
     h200: verified
     b200: verified
+    arc_pro_b60: verified
+    arc_pro_b70: verified
 
 model:
   model_id: "nvidia/NVIDIA-Nemotron-3-Nano-4B-BF16"
@@ -72,7 +74,11 @@ kv_offload_support:
   offloading_cpu: verified
   offloading_fs: verified
 
-hardware_overrides: {}
+hardware_overrides:
+  xpu:
+    extra_args:
+      - "--enforce-eager"
+
 strategy_overrides: {}
 
 guide: |
@@ -84,7 +90,7 @@ guide: |
 
   ## Prerequisites
 
-  - Hardware: 1x H100/H200/B200, DGX Spark, or Jetson Thor
+  - Hardware: 1x H100/H200/B200, DGX Spark, Jetson Thor, or Intel Arc Pro B60/B70
   - vLLM >= 0.11.2 (0.12.0 recommended)
   - Docker with NVIDIA Container Toolkit (recommended)
 
@@ -100,6 +106,12 @@ guide: |
   docker pull ghcr.io/nvidia-ai-iot/vllm:latest-jetson-thor
   ```
 
+  Intel XPU (B60 / B70):
+
+  ```bash
+  docker pull vllm/vllm-openai-xpu:latest
+  ```
+
   ## Launch commands
 
   BF16:
@@ -128,6 +140,47 @@ guide: |
     --tensor-parallel-size 1
   ```
 
+  ### Intel XPU (B60, B70)
+
+  Validated on 1× Intel Arc Pro B60 / B70 (B60 24 GB, B70 32 GB per card) with the
+  official vLLM XPU image `vllm/vllm-openai-xpu:latest`.
+
+  BF16:
+  ```bash
+  docker run --device /dev/dri \
+    -v /dev/dri/by-path:/dev/dri/by-path --shm-size=16g \
+    --privileged --ipc=host -p 8000:8000 \
+    -v ~/.cache/huggingface:/root/.cache/huggingface \
+    --entrypoint bash vllm/vllm-openai-xpu:latest \
+    -c "source /opt/intel/oneapi/setvars.sh && exec vllm serve nvidia/NVIDIA-Nemotron-3-Nano-4B-BF16 \
+    --trust-remote-code \
+    --async-scheduling \
+    --kv-cache-dtype auto \
+    --tensor-parallel-size 1 \
+    --enforce-eager \
+    --enable-auto-tool-choice \
+    --tool-call-parser qwen3_coder \
+    --reasoning-parser-plugin nano_v3_reasoning_parser.py \
+    --reasoning-parser nano_v3 \
+    --max-model-len 8192"
+  ```
+
+  FP8:
+  ```bash
+  docker run --device /dev/dri \
+    -v /dev/dri/by-path:/dev/dri/by-path --shm-size=16g \
+    --privileged --ipc=host -p 8000:8000 \
+    -v ~/.cache/huggingface:/root/.cache/huggingface \
+    --entrypoint bash vllm/vllm-openai-xpu:latest \
+    -c "source /opt/intel/oneapi/setvars.sh && exec vllm serve nvidia/NVIDIA-Nemotron-3-Nano-4B-FP8 \
+    --trust-remote-code \
+    --async-scheduling \
+    --kv-cache-dtype fp8 \
+    --tensor-parallel-size 1 \
+    --enforce-eager \
+    --max-model-len 8192"
+  ```
+
   ## Benchmarking
 
   ```bash