Merge branch 'master' into master

Author: lawrencedsz9

.github/copilot-instructions.md

@@ -99,6 +99,7 @@ Before posting any comment, apply this gate:
 - Do not review vendored/third-party sources under `thirdparty/` unless the PR explicitly modifies integration or patch logic.
 - Do not enforce component-specific runtime behavior rules on docs-only PRs.
 - For generated or auto-updated files (for example stubs/version bumps), comment only if there is clear breakage risk.
+- Python stub files (`.pyi`) under `src/bindings/python/` are auto-generated by a scheduled GitHub Actions workflow (`update_pyapi_stubs.yml`) using `pybind11-stubgen`. Do not request manual fixes to stub files.
 
 ## Component-Specific Expectations
 

.github/workflows/linux_conditional_compilation.yml

@@ -192,12 +192,14 @@ jobs:
             -DENABLE_DEBUG_CAPS=ON \
             -DCMAKE_C_COMPILER_LAUNCHER=${{ env.CMAKE_C_COMPILER_LAUNCHER }} \
             -DCMAKE_CXX_COMPILER_LAUNCHER=${{ env.CMAKE_CXX_COMPILER_LAUNCHER }} \
+            -DCMAKE_JOB_POOLS:STRING=link_jobs=8 \
+            -DCMAKE_JOB_POOL_LINK:STRING=link_jobs \
             -S ${OPENVINO_REPO} \
             -B ${BUILD_DIR}
 
       - name: Cmake build - CC COLLECT
         run: |
-          cmake --build ${BUILD_DIR} --parallel 8 --config ${{ env.CMAKE_BUILD_TYPE }}
+          cmake --build ${BUILD_DIR} --parallel $(nproc) --config ${{ env.CMAKE_BUILD_TYPE }}
           cmake --build ${BUILD_DIR} --parallel $(nproc) --config ${{ env.CMAKE_BUILD_TYPE }} --target sea_itt_lib
 
       - name: Show sccache stats
@@ -289,7 +291,7 @@ jobs:
     defaults:
       run:
         shell: bash
-    runs-on: aks-linux-16-cores-32gb
+    runs-on: aks-linux-8-cores-16gb
     container:
       image: ${{ fromJSON(needs.docker.outputs.images).ov_build.ubuntu_22_04_x64_cc }}
       volumes:

docs/articles_en/about-openvino/performance-benchmarks.rst

@@ -158,7 +158,7 @@ For a listing of all platforms and configurations used for testing, refer to the
 **Disclaimers**
 
 * System configurations used for Intel® Distribution of OpenVINO™ toolkit performance results
-  are based on release 2025.4, as of December 1st, 2025.
+  are based on release 2026.0, as of February 25, 2026.
 
 * OpenVINO Model Server performance results are based on release 2025.3, as of September 3rd, 2025.
 

docs/articles_en/about-openvino/performance-benchmarks/generative-ai-performance.rst

@@ -3,20 +3,31 @@ Most Efficient Large Language Models for AI PC
 
 This page is regularly updated to help you identify the best-performing LLMs on the
 Intel® Core™ Ultra processor family and AI PCs.
-The current data is as of OpenVINO 2025.3, 3 September 2025.
+The current data is as of OpenVINO 2026.0, 26 February 2026.
 
 The tables below list the key performance indicators for inference on built-in GPUs.
 
 
 
 .. tab-set::
 
-   .. tab-item:: 7-255H
+   .. tab-item:: 5-338H
 
       .. data-table::
          :class: modeldata stripe
          :name: supportedModelsTable_V1
          :header-rows: 1
+         :file:  ../../_static/benchmarks_files/llm_models_5-338H.csv
+         :data-column-hidden: [3,4,6]
+         :data-order: [[ 0, "asc" ]]
+         :data-page-length: 10
+
+   .. tab-item:: 7-255H
+
+      .. data-table::
+         :class: modeldata stripe
+         :name: supportedModelsTable_V2
+         :header-rows: 1
          :file:  ../../_static/benchmarks_files/llm_models_7-255H.csv
          :data-column-hidden: [3,4,6]
          :data-order: [[ 0, "asc" ]]
@@ -26,7 +37,7 @@ The tables below list the key performance indicators for inference on built-in G
 
       .. data-table::
          :class: modeldata stripe
-         :name: supportedModelsTable_V2
+         :name: supportedModelsTable_V3
          :header-rows: 1
          :file:  ../../_static/benchmarks_files/llm_models_7-258V.csv
          :data-column-hidden: [3,4,6]
@@ -36,7 +47,7 @@ The tables below list the key performance indicators for inference on built-in G
 
       .. data-table::
          :class: modeldata stripe
-         :name: supportedModelsTable_V3
+         :name: supportedModelsTable_V4
          :header-rows: 1
          :file:  ../../_static/benchmarks_files/llm_models_7-165H.csv
          :data-column-hidden: [3,4,6]
@@ -64,6 +75,13 @@ The tables below list the key performance indicators for inference on built-in G
 
          :material-regular:`download;1.5em` Get system descriptions [PDF]
 
+      .. button-link:: ../../_static/benchmarks_files/llm_models_5-338H.csv 
+         :color: primary
+         :outline:
+         :expand:
+
+         :material-regular:`download;1.5em` Get the 5-338H table [PDF]
+
       .. button-link:: ../../_static/benchmarks_files/llm_models_7-258V.csv 
          :color: primary
          :outline:

docs/articles_en/about-openvino/performance-benchmarks/model-accuracy-int8-fp32.rst

@@ -11,7 +11,7 @@ the table for more information.
 * A - Intel® Core™ Ultra 9-185H (AVX2), INT8 and FP32
 * B - Intel® Xeon® 6338, (VNNI), INT8 and FP32
 * C - Intel® Xeon 6972P (VNNI, AMX), INT8, BF16, FP32
-* D - Intel® Arc-B580, INT8 and FP16
+* D - Intel® Arc-B60, INT8 and FP16
 
 
 .. list-table:: Model Accuracy for INT8
@@ -27,10 +27,10 @@ the table for more information.
    * - bert-base-cased
      - SST-2_bert_cased_padded
      - spearman@cosine
-     - 2.57%
-     - 2.65%
-     - 2.95%
-     - 2.63%
+     - 2.60%
+     - 2.70%
+     - 3.00%
+     - 2.60%
    * - Detectron-V2
      - COCO2017_detection_91cl_bkgr
      - coco_orig_precision
@@ -43,22 +43,22 @@ the table for more information.
      - accuracy @ top1
      - -0.91%
      - -0.93%
-     - -0.91%
-     - -1.03%
+     - -1.01%
+     - -1.01%
    * - resnet-50
      - ImageNet2012
      - accuracy @ top1
-     - -0.12%
-     - -0.12%
-     - -0.15%
-     - -0.15%
+     - 0.73%
+     - 0.73%
+     - 0.73%
+     - 0.73%
    * - ssd-resnet34-1200
      - COCO2017_detection_80cl_bkgr
      - map
-     - 0.00%
-     - 0.00%
-     - 0.03%
-     - 0.07%
+     - 0.02%
+     - 0.02%
+     - 0.02%
+     - 0.02%
    * - yolo_v11
      - COCO2017_detection_80cl
      - AP@0.5:0.05:0.95
@@ -93,31 +93,31 @@ the table for more information.
    * - mobilenet-v2
      - ImageNet2012
      - accuracy @ top1
-     - 0.00%
-     - 0.00%
-     - 0.02%
-     - 0.02%
+     - -0.01%
+     - -0.01%
+     - -0.01%
+     - -0.01%
    * - resnet-50
      - ImageNet2012
      - accuracy @ top1
      - 0.00%
      - 0.00%
      - 0.00%
-     - -0.01%
+     - 0.00%
    * - ssd-resnet34-1200
      - COCO2017_detection_80cl_bkgr
      - map
      - 0.02%
      - 0.02%
      - 0.02%
-     - -0.23%
+     - 0.02%
    * - yolo_v11
      - COCO2017_detection_80cl
      - AP@0.5:0.05:0.95
-     - 0.03%
-     - -2.21%
-     - -2.21%
-     - -2.21%
+     - 0.00%
+     - -2.18%
+     - -2.18%
+     - -2.18%
 .. list-table:: Model Accuracy for AMX-FP16, AMX-INT4, Arc-FP16 and Arc-INT4 (Arc™ B-series)
    :header-rows: 1
 
@@ -132,79 +132,93 @@ the table for more information.
      - Data Default WWB
      - Similarity
      - 98.1%
-     - 94.4%
-     - 99.5%
+     - 94.1%
+     - 99.6%
      - 94.0%
    * - DeepSeek-R1-Distill-Qwen-1.5B
      - Data Default WWB
      - Similarity
      - 96.5%
      - 92.4%
      - 99.7%
-     - 92.3%
+     - 92.7%
    * - Gemma-3-4B-it
      - Data Default WWB
      - Similarity
-     - 92.0%
+     - 92.2%
      - 83.9%
      - 
-     - 84.9%
+     - 92.9%
+   * - GPT-OSS-20B
+     - Data Default WWB
+     - Similarity
+     - 94.9%
+     - 92.2%
+     - 
+     - 92.9%
    * - Llama-2-7B-chat
      - Data Default WWB
      - Similarity
      - 99.3%
-     - 93.4%
-     - 99.8%
-     - 93.4%
+     - 93.3%
+     - 99.6%
+     - 93.3%
    * - Llama-3-8B
      - Data Default WWB
      - Similarity
      - 98.8%
-     - 94.3%
-     - 99.7%
-     - 94.5%
+     - 94.7%
+     - 99.9%
+     - 94.4%
    * - Llama-3.2-3b-instruct
      - Data Default WWB
      - Similarity
-     - 97.9%
-     - 94.2%
-     - 99.7%
-     - 94.1%
-   * - Phi4-mini-instruct
+     - 98.3%
+     - 94.8%
+     - 99.9%
+     - 94.3%
+   * - MiniCPM-V-2.6
      - Data Default WWB
      - Similarity
+     - 90.6%
+     - 90.1%
+     - 88.1%
      - 89.1%
+   * - Phi4-mini-instruct
+     - Data Default WWB
+     - Similarity
+     - 95.1%
+     - 92.5%
+     - 99.1%
      - 92.1%
-     - 99.5%
-     - 92.4%
-   * - Qwen2-VL-7B
+   * - Qwen2.5-VL-7B
      - Data Default WWB
      - Similarity
-     - 97.5%
-     - 88.1%
+     - 93.7%
+     - 90.7%
      - 99.8%
-     - 91.4%
+     - 89.9%
    * - Qwen3-8B
      - Data Default WWB
      - Similarity
-     - 97.8%
-     - 92.3%
-     - 
-     - 93.0%
+     - 97.9%
+     - 93.6%
+     - 99.8%
+     - 92.8%
    * - Flux.1-schnell
      - Data Default WWB
      - Similarity
      - 95.4%
      - 96.1%
      - 
-     - 92.1%
+     - 95.1%
    * - Stable-Diffusion-V1-5
      - Data Default WWB
      - Similarity
-     - 96.3%
-     - 93.3%
+     - 96.7%
+     - 95.5%
      - 99.5%
-     - 93.7%
+     - 92.1%
 
 Notes: For all accuracy metrics a "-", (minus sign), indicates an accuracy drop.
 The Similarity metric is the distance from "perfect" and as such always positive. 

docs/articles_en/about-openvino/performance-benchmarks/performance-benchmarks-faq.rst

@@ -59,6 +59,10 @@ Performance Information F.A.Q.
         - Hugginface
         - Text-To-Text Decoder-only
         - 128K
+      * - `GPT-OSS-20B <https://huggingface.co/openai/gpt-oss-20b>`__
+        - Hugginface
+        - Reasoning, agentic and developer use cases.
+        - 128K
       * - `Llama-2-7b-chat <https://www.llama.com/>`__
         - Meta AI
         - Auto regressive language
@@ -71,13 +75,21 @@ Performance Information F.A.Q.
         - Meta AI
         - Auto regressive language
         - 128K
+      * - `MiniCPM-v-2_6 <https://huggingface.co/openbmb/MiniCPM-V-2_6>`__
+        - OpenBMB
+        - Multimodal, image-text to text
+        - 32K
+      * - `Mistral-7B-Instruct-V0.3 <https://huggingface.co/mistralai/Mistral-7B-Instruct-v0.3>`__
+        - Mistral AI
+        - Auto regressive language
+        - 32K
       * - `Phi4-mini-Instruct <https://huggingface.co/microsoft/Phi-4-mini-instruct>`__
-        - Huggingface
+        - Hugginface
         - Auto regressive language
         - 4096
-      * - `Qwen-2-VL-7B-instruct <https://huggingface.co/Qwen/Qwen2-VL-7B-instruct>`__
+      * - `Qwen-2.5-VL-7B-instruct <https://huggingface.co/Qwen/Qwen2.5-VL-7B-instruct>`__
         - Huggingface
-        - Auto regressive language
+        - Auto regressive language, Agentic
         - 128K
       * - `Qwen-3-8B <https://huggingface.co/Qwen/Qwen3-8B>`__
         - Huggingface