What changed in OpenVINO from 2025 to 2026

9 releases numbered after 2025.4.1 up to and including 2026.4.1, stable releases only. 2025.4.1 and 2026.4.1 are the newest stable releases of 2025 and 2026 we track; this page follows them as new ones ship.

165 changes across 9 releases · 13 landed on more than one version

Added 66

2026.4.1

  • Added causal-attention and partial rotary-embedding support to GroupQueryAttention in the ONNX frontend
  • Enabled f8e4m3 Gather for the windowed KV-cache

2026.4.0

  • New CPU support for the Gemma-3n model
  • New CPU and GPU support for Kokoro-82M, Qwen3-VL-4B with EAGLE-3, Qwen3-ASR, Muse Glimmer 30B, Qwen3.8 27B, Gemma 4 12B, Hy-MT2-1.8B, DeepSeek OCR-2, and Granite 4.0 H Micro models
  • New NPU support for FLUX.2-Klein 4B and Kokoro-82M models
  • Early-release CPU and GPU support for Qwen-Image, Z-Image-Turbo, Granite 4.0 H Tiny, Fun-ASR-Nano, LFM2.5-8B-A1B, MiniCPM5-2B, RF-DETR, BGE Reranker-V2-M3, and BGE M3 models
  • Multi-Token Prediction (MTP) speculative decoding in OpenVINO GenAI for Gemma 4, Qwen3.5, and Qwen3.6 on CPUs and GPUs
  • Preview DFlash acceleration in OpenVINO GenAI for Qwen models on GPUs
  • Preview visual-token support in OpenVINO GenAI to reduce latency on Intel Core Ultra Series 3 processors
  • Tree Drafting (Top-K) for EAGLE-3 in OpenVINO GenAI for higher VLM pipeline throughput compared to Chain Drafting (Top-1)
  • ITT profiling support for the NPU, allowing Intel VTune Profiler to analyze CPU, GPU, and NPU execution
  • ASRPipeline support in Node.js in OpenVINO GenAI, with streaming and performance metrics for models such as Whisper and Qwen3-ASR
  • Preview bounded dynamic-shape support on NPUs for vision workloads such as image super-resolution, validated with the ESPCN model
  • Preview idle model management in OpenVINO Model Server that unloads unused models to reduce memory usage
  • OpenVINO Model Server support for agentic models Muse Glimmer 30B and Qwen3.8 27B

2026.3.1

  • Functional enablement of the Muse-Glimmer-30B model, with model card and IR available.
  • Functional enablement of the Qwen3.8-27B model, with model card and IR available.

2026.3.0

  • Support for SmolLM3-3B, LFM2-1.2B, and LFM2.5-1.2B models on CPU, GPU, and NPU
  • Support for Harrier OSS-v1-0.6B, Qwen3-8B with EAGLE-3, MiniCPM5-1B, and FLUX.2-klein-4B models on CPU and GPU
  • YOLO26 model support extended to GPU and NPU
  • Early-release support for Qwen3-ASR, Qwen3-Omni, Gemma-3n, Qwen3-VL-Embedding-8B, and Kokoro-82M models on CPU and GPU
  • Support for Hugging Face Transformers v5.5
  • EAGLE-3 speculative decoding pipeline extended to LLMs and VLMs in OpenVINO GenAI, with Top-K sampling added for additional token-generation speedups on CPUs, GPUs, and NPUs
  • Lazy weight loading for IR and ONNX models that automatically selects an optimal loading and compilation path to minimize peak memory usage during model initialization
  • FP8 quantization for ONNX models in NNCF
  • Omni pipeline for multimodal workloads, ASR pipeline for speech recognition, and Embedding pipeline for multimodal embedding generation in OpenVINO GenAI
  • Support for Intel Xeon 6+ processors (formerly codenamed Clearwater Forest)
  • MoE offloading to disk, allowing 30B MoE models such as Qwen3-30B-A3B to run on devices with 16 GB of memory at acceptable tokens-per-second rates
  • Standard v1/chat/completions support in OpenVINO Model Server
  • Tool-parser support extended to MiniCPM5-1B and LFM2.5 in OpenVINO Model Server

2026.2.0

  • Support for new models Gemma 4 E2B and Gemma 4 E4B
  • Support for Qwen3-Coder-Next, Qwen3.5, Qwen3.6, Trinity-mini, LFM2-24B-A2B, LFM2-8B-A1B, and LFM2.5-350M models on CPUs and GPUs
  • Support for YOLO26 model on CPUs
  • Support for Gemma 4 31B and Gemma 4 26B-A4B models on GPUs
  • Scaled Dot-Product Attention (SDPA) path support for LFM2 models
  • Support for Hugging Face Transformers v5.0
  • OpenVINO GenAI extensions property for loading custom extension libraries and registering unsupported operations
  • INT4 KV-cache compression enabled for GPUs
  • Preview: by-channel INT8 KV-cache quantization for enhanced XAttention accuracy on CPUs and GPUs
  • OpenVINO GenAI JavaScript API includes a Text-to-Speech pipeline and VLM samples for browser and Node.js
  • Prompt Lookup Decoding extended to vision-language pipelines
  • OpenVINO Model Server tool-calling support for Qwen 3.5 and 3.6 models
  • OpenVINO Model Server streaming transcription support for speech-to-text

2026.1.0

  • Support for the Qwen3 VL model on CPUs and GPUs
  • Support for the GPT-OSS 120B model on CPUs
  • Preview of the OpenVINO backend for llama.cpp, enabling optimized inference on Intel CPUs, GPUs, and NPUs
  • Unified VLM chatbot notebook with video file support and interactive model switching across Qwen3-VL, Qwen2.5-VL, and LLaVa-NeXT-Video
  • TaylorSeer Lite caching in OpenVINO GenAI for image and video generation, accelerating diffusion-transformer inference across Flux, SD3, and LTX-Video pipelines
  • Dynamic LoRA support in OpenVINO GenAI for Qwen3-VL and VL models with LLM, allowing adapters to be swapped at runtime without reloading the base model
  • Preview of the release-weights API for ov::Model, reclaiming memory during model compilation on NPUs
  • Support for Intel Core Series 3 processors (formerly Wildcat Lake) and Intel Arc Pro B70 Graphics with 32GB memory for single-GPU inference on 20-30B parameter LLMs
  • Prompt Lookup Decoding extended to vision-language pipelines on Intel CPUs and GPUs
  • WhisperPipeline for Node.js in OpenVINO GenAI via its NPM package, with word-level audio-to-text transcription
  • Image inpainting and outpainting via the /image endpoint in OpenVINO Model Server

2026.0.0

  • New models supported on CPUs and GPUs: GPT-OSS-20B, MiniCPM-V-4_5-8B, and MiniCPM-o-2.6
  • New models supported on NPUs: MiniCPM-o-2.6, plus NPU support for Qwen2.5-1.5B-Instruct, Qwen3-Embedding-0.6B, and Qwen-2.5-coder-0.5B
  • Word-level timestamp functionality in the OpenVINO GenAI Whisper Pipeline on CPUs, GPUs, and NPUs
  • Phi-3-mini FastDraft model available on Hugging Face to accelerate LLM inference on NPUs
  • Int4 data-aware weight compression for 3D MatMuls in the Neural Network Compression Framework, enabling MoE LLMs to run with reduced memory and bandwidth
  • Preview support for per-layer and per-group Look-Up Tables (LUT) for FP8-4BLUT quantization in the Neural Network Compression Framework
  • Preview VLM pipeline support in OpenVINO GenAI to enhance Agentic AI framework integration
  • Speculative decoding support for NPUs in OpenVINO GenAI
  • Preview NPU compiler integration with the NPU plugin enabling ahead-of-time and on-device compilation without relying on OEM driver updates
  • Enhanced audio endpoint support, agentic continuous batching, and concurrent runs in OpenVINO Model Server
  • Updated Node.js bindings compatible with Node.js v22 LTS, not supporting CentOS 7

2026.0.1

  • Preview NPU compiler integration with the NPU plugin enabling ahead-of-time and on-device compilation without relying on OEM driver updates, enabled by default in this release package
Changed 24

2026.4.1

  • Extended MatMulNBits with bfloat16 support and reordered weight, scale, and zero-point layouts for correct decoding of more quantized models
  • The OV_CPU_MAX_ISA instruction-set cap now works in release builds (previously debug-only) and is modeled as a feature bitmask

2026.4.0

  • Xe3 integrated graphics optimizations in OpenVINO GenAI improve inference performance for Gemma 4 models with long-context inputs on Intel Core Ultra Series 3 processors
  • PostponedConstant constructor signature changed from Callable[[Tensor], None] to Callable[[], Tensor]also in2026.0.0
  • CPU plugin now requires AVX2 instruction set support as a minimum system requirement; SSE is no longer supportedalso in2026.3.02026.2.02026.1.02026.0.0

2026.3.0

  • OpenVINO Model Server unifies REST API endpoints and reduces command complexity for model deployment
  • OpenVINO Model Server improves stability, performance, and accuracy for LLMs such as Qwen3.5/3.6 with linear attention
  • OpenVINO builds migrated from RHEL 8 to RHEL 9also in2026.2.02026.1.0
  • PostponedConstant constructor signature updated from Callable[[Tensor], None] to Callable[[], Tensor]also in2026.2.02026.1.0

2026.2.0

  • GPT-OSS-120B model support extended to GPUs
  • OpenVINO GenAI significantly reduces model loading times on GPU when using cache blobs
  • Optimized IR read mode with independently managed constant buffers to reduce peak memory usage, with Linux support added
  • OpenVINO GenAI has a smaller runtime footprint after eliminating ICU DLL dependencies from tokenization
  • manylinux2014 upgraded to manylinux_2_28, and CentOS 7 is no longer supported due to glibc incompatibility
  • Updated Node.js bindings compatible with Node.js v22 LTS; these bindings do not support CentOS 7also in2026.1.0

2026.1.0

  • LTX-Video generation on GPU achieves end-to-end acceleration through fusion of RMSNorm and RoPE operators
  • OpenVINO GenAI runtime footprint reduced by eliminating ICU DLL dependencies from tokenization, lowering memory usage and startup time
  • OpenVINO Model Server enhanced support for Qwen3-MOE and GPT-OSS-20b models with improved performance, accuracy, and concurrent request handling
  • manylinux2014 upgraded to manylinux_2_28, dropping CentOS 7 support due to glibc incompatibility
  • APT and YUM repositories restructured to remove year-based subdirectories; the legacy structure is unavailable starting 2026.0

2026.0.0

  • Builds migrated from RHEL 8 to RHEL 9
  • manylinux2014 upgraded to manylinux_2_28, which drops CentOS 7 support due to glibc incompatibility
  • APT and YUM repositories restructured to no longer use year-based subdirectories; the legacy structure is unavailable starting 2026.0
  • Major internal refactoring of the graph iteration mechanism; the legacy path can be enabled with ONNX_ITERATOR=0 and is deprecated
Fixed 15

2026.4.1

  • Fixed an out-of-bounds ScatterUpdate index in the sliding-window KV-cache path that could cause memory corruption or a driver crash on GPUs
  • Fixed incorrect Gather shape inference for f8e4m3 and other low-precision data types
  • Fixed wrong or corrupted results on GPU when using remote zero-copy output tensors with a transpose at the model output under dynamic shapes
  • Fixed GRU sequence (RNN) models producing incorrect results on GPU with dynamic shapes
  • Corrected blocked-format reorder offsets on GPU
  • Fixed recursive layout and format propagation through Reshape nodes during GPU graph compilation
  • Fixed unexpected application crashes on Intel Core Ultra Series 3 processors on Windows with certain Intel NPU PV drivers (32.0.100.4300, 32.0.100.4509, and 32.0.100.4512)
  • Aligned JIT kernel selection with the OV_CPU_MAX_ISA cap so operators produce correct results when the ISA is capped

2026.3.1

  • Property ENABLE_CPU_PINNING is now supported on dual-socket Intel Xeon 6+ processors (formerly codenamed Clearwater Forest) Windows Servers.
  • Chat template updated for Gemma 4 models so it can be parsed by the Minja templating engine.
  • Accuracy issue for the Gemma 4 26B A4B model when the input prompt is long.
  • Accuracy issue for the Gemma 4 31B model when an image is included in the input.
  • Accuracy issue for lfm2-1.2b and lfm2.5-1.2b-instruct models on the NPU device target.

2026.2.1

  • YOLO26 no longer fails to compile on GPU (issue ID 187077)
  • NPU plugin no longer violates app/inference priorities when using a shared L0 command queue
Removed 39

2026.4.0

  • The deprecated openvino.runtime namespace, which should be replaced by the openvino namespace
  • The deprecated openvino.Type.undefined, which should be replaced by openvino.Type.dynamic
  • Support for Debian 10also in2026.3.0
  • Deprecated OpenVINO GenAI predefined generation configsalso in2026.3.0
  • Deprecated OpenVINO GenAI support for the whisper stateless decoder model, which should be replaced by a stateful model
  • Deprecated OpenVINO GenAI StreamerBase put method, bool return type for callbacks, and ChunkStreamer classalso in2026.3.0
  • NNCF optimization methods for TensorFlow models and the TensorFlow backendalso in2026.3.02026.0.0
  • Experimental NNCF methods NAS, Structural Pruning, AutoML, Knowledge Distillation, Mixed-Precision Quantization, and Movement Sparsity
  • TensorFlow Serving (TFS) API support in OpenVINO Model Server 2026.3, with KServe API recommended instead
  • Stateful Models support in OpenVINO Model Server 2026.3

2026.3.0

  • Deprecated openvino.runtime namespace; use the openvino namespace directly
  • Deprecated openvino.Type.undefined; use openvino.Type.dynamic instead
  • Deprecated OpenVINO GenAI support for whisper stateless decoder model; use a stateful model instead
  • NNCF create_compressed_model() method; use nncf.prune() for unstructured pruning and nncf.quantize() for INT8 quantizationalso in2026.0.0
  • Experimental NNCF methods: NAS, Structural Pruning, AutoML, Knowledge Distillation, Mixed-Precision Quantization, and Movement Sparsity
  • TensorFlow Serving (TFS) API in OpenVINO Model Server 2026.3; KServe API is recommended instead

2026.2.0

  • Removed the deprecated openvino.runtime namespace; use the openvino namespace directly
  • Removed the deprecated openvino.Type.undefined; use openvino.Type.dynamic instead
  • Removed the deprecated OpenVINO GenAI predefined generation configs
  • Removed OpenVINO GenAI support for the whisper stateless decoder model; use a stateful model
  • Removed the deprecated OpenVINO GenAI StreamerBase put method, bool return type for callbacks, and ChunkStreamer class
  • NNCF optimization methods for TensorFlow models and the TensorFlow backend removed in 2026
  • Removed experimental NNCF methods: NAS, Structural Pruning, AutoML, Knowledge Distillation, Mixed-Precision Quantization, and Movement Sparsity

2026.1.0

  • The deprecated openvino.runtime namespace has been removed; use the openvino namespace directly
  • The deprecated openvino.Type.undefined has been removed; use openvino.Type.dynamic instead
  • The deprecated OpenVINO GenAI predefined generation configs were removed
  • Support for the whisper stateless decoder model in OpenVINO GenAI has been removed; use a stateful model
  • The deprecated StreamerBase put method, bool return type for callbacks, and ChunkStreamer class were removed from OpenVINO GenAI
  • NNCF optimization methods for TensorFlow models and the TensorFlow backend were removed in 2026
  • Experimental NNCF methods NAS, Structural Pruning, AutoML, Knowledge Distillation, Mixed-Precision Quantization, and Movement Sparsity were removed
  • MacOS x86 is no longer supported
  • OpenCV binaries removed from Docker imagesalso in2026.0.0

2026.0.0

  • The deprecated openvino.runtime namespace; use the openvino namespace directly
  • The deprecated openvino.Type.undefined; use openvino.Type.dynamic instead
  • The deprecated OpenVINO GenAI predefined generation configs
  • Deprecated OpenVINO GenAI support for the whisper stateless decoder model; use a stateful model instead
  • The deprecated OpenVINO GenAI StreamerBase put method, bool return type for callbacks, and ChunkStreamer class
  • The experimental NNCF methods NAS, Structural Pruning, AutoML, Knowledge Distillation, Mixed-Precision Quantization, and Movement Sparsity
  • MacOS x86 support
Deprecated 17

2026.4.0

  • NNCF create_compressed_model() method, to be replaced by nncf.prune() for unstructured pruning and nncf.quantize() for INT8 quantization
  • Support for Python 3.10 will be discontinued in OpenVINO 2026.5 due to end-of-life status
  • auto shape and auto batch size runtime model reshaping, to be replaced by dynamic shape models
  • Legacy graph iteration path enabled via ONNX_ITERATOR=0, to be removed in a future release
  • Modifying runtime information (RTMap) through read-only ConstOutput objects in the Python API, to be removed in 2027.0

2026.3.0

  • auto shape and auto batch size (runtime model reshaping); dynamic shape models are recommended instead
  • Legacy ONNX graph iteration path, enabled via ONNX_ITERATOR=0, will be removed in future releases
  • Dedicated OpenVINO operator for Kubernetes and OpenShift in favor of the KServe operator
  • Stateful models support in OpenVINO Model Server will be deprecated

2026.2.0

  • NNCF create_compressed_model() method deprecated and removed in 2026; use nncf.prune() or nncf.quantize() instead
  • Support for Ubuntu 20.04 discontinued due to end of standard supportalso in2026.0.0
  • auto shape and auto batch size (runtime model reshaping) will be removed in the future; dynamic shape models are recommended
  • Legacy ONNX graph iteration path (enabled via ONNX_ITERATOR=0) is deprecated and will be removed in future releases
  • Dedicated OpenVINO operator for Kubernetes and OpenShift deprecated in favor of the KServe operator

2026.1.0

  • NNCF create_compressed_model() method is deprecated and removed in 2026; use nncf.prune() for unstructured pruning and nncf.quantize() for INT8 quantization
  • Support for Ubuntu 20.04 has been discontinued due to end of standard support
  • auto shape and auto batch size (runtime model reshaping) will be removed in the future; dynamic shape models are recommended insteadalso in2026.0.0
Security 4

2026.4.1

  • Hardened compiled-blob import against malformed or forged blobs by fixing a possible integer overflow in metadata parsing, bounding recursion in the compatibility-string parser, and rejecting forged blobs on import
  • Added index validation to ScatterElementsUpdate
  • Added divide-by-zero and overflow guards to Col2Im shape inference
  • Added axes validation preventing a potential out-of-bounds write in the ONNX Pad operator

Original release notes, newest first

The list above is our reading of these notes; the originals from Intel are here, one fold per release.

2026.4.1
Summary of major features and improvements
  • Improved GroupQueryAttention support for large language models served through the ONNX frontend:
    • Added causal-attention and partial rotary-embedding support to GroupQueryAttention, broadening coverage of modern LLMs.
    • Extended MatMulNBits with bfloat16 support and reordered weight/scale/zero-point layouts for correct decoding of more quantized models.
    • Fixed correctness issues in the sliding-window KV-cache path that could produce wrong results:
      • an out-of-bounds ScatterUpdate index that could cause memory corruption or a driver crash on GPUs,
      • incorrect Gather shape inference for f8e4m3 and other low-precision data types (f8e4m3 Gather is now enabled for the windowed KV-cache).
  • Fixed issues:
    • GPU:
      • Fixed wrong or corrupted results when using remote (zero-copy) output tensors with a transpose at the model output under dynamic shapes.
      • Fixed GRU sequence (RNN) models producing incorrect results with dynamic shapes, and corrected blocked-format reorder offsets.
      • Fixed recursive layout/format propagation through Reshape nodes during GPU graph compilation.
    • NPU:
      • ID 233822: Fixed unexpected application crashes on Intel® Core™ Ultra Series 3 processors on Windows that could occur with certain Intel® NPU PV drivers (32.0.100.4300, 32.0.100.4509, and 32.0.100.4512). It is recommended to migrate to OV 2026.4.1 in case the issue is observed with OV releases 2026.2, 2026.3, or 2026.4.
      • Hardened compiled-blob import against malformed or forged blobs: fixed a possible integer overflow in metadata parsing, bounded recursion in the compatibility-string parser, and rejected forged blobs on import.
    • CPU:
      • The OV_CPU_MAX_ISA instruction-set cap now works in release builds (previously debug-only) and is modeled as a feature bitmask.
      • Aligned JIT kernel selection with the OV_CPU_MAX_ISA cap so operators produce correct results when the ISA is capped.
    • Core / Frontends:
      • Hardened operators against malformed inputs: added index validation to ScatterElementsUpdate, divide-by-zero and overflow guards to Col2Im shape inference, and axes validation preventing a potential out-of-bounds write in the ONNX Pad operator.

You can find the OpenVINO™ toolkit 2026.4.1 release here:

Release documentation is available here: https://docs.openvino.ai/2026

Release Notes are available here: https://docs.openvino.ai/2026/about-openvino/release-notes-openvino.html

*Other names and brands may be claimed as the property of others.

View originalPermalink

2026.4.0
Summary of major features and improvements  
  • More GenAI coverage and framework integrations to minimize code changes
    • New models supported:
      • On CPU: Gemma-3n
      • On CPU, GPU: Kokoro-82M, Qwen3-VL-4B with EAGLE-3, Qwen3-ASR, Muse Glimmer 30B, Qwen3.8 27B, Gemma 4 12B, Hy-MT2-1.8B, DeepSeek OCR-2, and Granite 4.0 H Micro
      • On NPUs: FLUX.2-Klein 4B and Kokoro-82M
    • Additional CPU and GPU-enabled models available as early releases: Qwen-Image, Z-Image-Turbo, Granite 4.0 H Tiny, Fun-ASR-Nano, LFM2.5-8B-A1B, MiniCPM5-2B, RF-DETR, BGE Reranker-V2-M3, and BGE M3
  • Broader LLM model support and more model compression techniques
    • OpenVINO™ GenAI adds Multi-Token Prediction (MTP) speculative decoding for Gemma 4, Qwen3.5, and Qwen3.6 on CPUs and GPUs, enabling higher throughput and lower latency without sacrificing accuracy.
    • Preview: OpenVINO™ GenAI introduces DFlash acceleration for Qwen models on GPUs, and visual-token support to reduce latency and speed up GenAI pipelines on Intel® Core™ Ultra Series 3 processors.
    • With Tree Drafting (Top-K) for EAGLE-3 now supported in OpenVINO™ GenAI, developers can unlock higher throughput on VLM pipelines compared to Chain Drafting (Top-1).
    • Xe3 integrated graphics optimizations in OpenVINO™ GenAI improve AI inference performance for Gemma 4 models processing long-context inputs on Intel® Core™ Ultra Series 3 processors.
    • OpenVINO extends Instrumentation and Tracing Technology (ITT) profiling support to the NPU, enabling developers to use Intel® VTune™ Profiler to analyze CPU, GPU, and NPU execution through one consistent toolchain.
  • More portability and performance to run AI at the edge, in the cloud, or locally.
    • OpenVINO™ GenAI introduces support for ASRPipeline in Node.js, enabling JavaScript developers to run automatic speech recognition (such as Whisper and Qwen3-ASR) with streaming and performance metrics using a pipeline API similar to C++ and Python.
    • Preview: Bounded dynamic-shape support on NPUs now extends to vision workloads such as image super-resolution. This capability has been validated with the ESPCN model.
    • Preview: OpenVINO™ Model Server now includes preview support for idle model management, which unloads models when they are not in use to reduce memory usage.
    • OpenVINO™ Model Server adds support for new agentic models such as Muse Glimmer 30B and Qwen3.8 27B.
Support Change and Deprecation Notices
  • Discontinued in 2026.0:

    • The deprecated openvino.runtime namespace has been removed. Please use the openvino namespace directly.
    • The deprecated openvino.Type.undefined has been removed. Please use openvino.Type.dynamic instead.
    • Support for Debian 10 has been discontinued due to the end of its standard support.
    • The PostponedConstant constructor signature has been updated for improved usability:
      • Old (removed): Callable[[Tensor], None]
      • New: Callable[[], Tensor]
    • The deprecated OpenVINO™ GenAI predefined generation configs were removed.
    • The deprecated OpenVINO GenAI support for whisper stateless decoder model has been removed. Please use a stateful model.
    • The deprecated OpenVINO GenAI StreamerBase put method, bool return type for callbacks, and ChunkStreamer class has been removed.
    • NNCF create_compressed_model() method is now deprecated and removed in 2026. Please use nncf.prune() method for unstructured pruning and nncf.quantize() for INT8 quantization.
    • NNCF optimization methods for TensorFlow models and TensorFlow backend in NNCF are deprecated and removed in 2026. It is recommended to use PyTorch analogous models for training-aware optimization methods and OpenVINO™ IR, PyTorch, and ONNX models for post-training optimization methods from NNCF.
    • The following experimental NNCF methods are deprecated and removed: NAS, Structural Pruning, AutoML, Knowledge Distillation, Mixed-Precision Quantization, Movement Sparsity.
    • CPU plugin now requires support for the AVX2 instruction set as a minimum system requirement. The SSE instruction set will no longer be supported.
    • Dropped support for the TensorFlow Serving (TFS) API in OpenVINO Model Server 2026.3. KServe API is recommended for classic model deployments.
    • Support for Stateful Models was removed in OpenVINO Model Server 2026.3. These capabilities were originally introduced for Kaldi audio models which are no longer applicable. Current audio model support relies on the OpenAI API and pipelines implemented with the OpenVINO™ GenAI library.
    • Support for Python 3.10 will be discontinued in OpenVINO 2026.5, due to its end-of-life (EOL) status.
  • Deprecated and to be removed in the future:

    • auto shape and auto batch size (reshaping a model in runtime) will be removed in the future. OpenVINO’s dynamic shape models are recommended instead.
    • Starting with 2026.0 release major internal refactoring of the graph iteration mechanism has been implemented for improved performance and maintainability. The legacy path can be enabled by setting the ONNX_ITERATOR=0 environment variable. This legacy path is deprecated and will be removed in future releases.
    • Modifying runtime information (RTMap) through read-only ConstOutput objects in the Python API is deprecated and will be removed in a 2027.0 release. Users should use non-const outputs when updating runtime metadata; attempts to modify RTMap through ConstOutput now emit a deprecation warning to ease migration.
    • OpenVINO Model Server:
      • The dedicated OpenVINO operator for Kubernetes and OpenShift is now deprecated in favor of the recommended KServe operator. The OpenVINO operator will remain functional in upcoming OpenVINO™ Model Server releases but will no longer be actively developed. Since KServe provides broader capabilities, no loss of functionality is expected. On the contrary, more functionalities will be accessible and migration between other serving solutions and OpenVINO Model Server will be much easier.
      • Directed Acyclic Graph Scheduler will be deprecated in favor of pipelines managed by MediaPipe scheduler and will be removed in 2026.3. That approach gives more flexibility, includes wider range of calculators and has support for using processing accelerators.
      • OpenVINO™ GenAI:
        • start_chat() / finish_chat() APIs are deprecated and will be removed in a future major release. Pass a ChatHistory object directly to generate() instead.

Acknowledgements

Thanks for contributions from the OpenVINO developer community: @ainergiz @Akihiro-Tabuchi @aobolensk @ash1shb1sht @da3dsoul @deepak0x @goyaladitya05 @hrushitfujitsu @jatinwadhwa921 @Lagmator22 @lxcxjxhx @mostafafaheem @NaiveTomcat @nuclearcat @odili @Passavee-Losripat @samwebster @vtanisha

You can find OpenVINO™ toolkit 2026.4 release here:

Release documentation is available here: https://docs.openvino.ai/2026 Release Notes are available here: https://docs.openvino.ai/2026/about-openvino/release-notes-openvino.html

View originalPermalink

2026.3.1
Summary of major features and improvements  

You can find OpenVINO™ toolkit 2026.3.1 release here:

Release documentation is available here: https://docs.openvino.ai/2026 Release Notes are available here: https://docs.openvino.ai/2026/about-openvino/release-notes-openvino.html

View originalPermalink

2026.3.0
Summary of major features and improvements  
  • More GenAI coverage and framework integrations to minimize code changes
    • New models supported:
      • On CPU, GPU, and NPU: SmolLM3-3B, LFM2-1.2B, LFM2.5-1.2B
      • On CPU and GPU: Harrier OSS-v1-0.6B, Qwen3-8B with EAGLE-3, MiniCPM5-1B, FLUX.2-klein-4B
      • Extended to GPU and NPU: YOLO26
      • Additional models available as early releases on CPU and GPU: Qwen3-ASR, Qwen3-Omni, Gemma-3n, Qwen3-VL-Embedding-8B, Kokoro-82M
    • Support for Hugging Face Transformers v5.5, ensuring compatibility with the latest model architectures on Hugging Face
  • Broader LLM model support and more model compression techniques
    • OpenVINO™ GenAI extends the EAGLE-3 speculative decoding pipeline to LLMs and VLMs, enhancing existing continuous batching and adding Top-K sampling to deliver additional token-generation speedups on CPUs, GPUs, and NPUs.
    • Lazy weight loading is enabled for IR and ONNX models to automatically select an optimal loading and compilation path, minimizing peak memory usage during model initialization.
    • Neural Network Compression Framework (NNCF) now supports FP8 quantization for ONNX models, helping developers realize FP8 performance, accuracy, and memory gains while expanding low-precision inference options for production deployments.
    • OpenVINO™ GenAI now supports three additional pipelines: Omni for multimodal workloads, ASR for speech recognition, and Embedding for multimodal embedding generation.
  • More portability and performance to run AI at the edge, in the cloud, or locally.
    • Introducing support for Intel® Xeon® 6+ processors (formerly codenamed Clearwater Forest)
    • MoE offloading to disk enabled, allowing 30B MoE models like Qwen3-30B-A3B to run even on devices with 16 GB of memory while maintaining acceptable tokens-per-second (TPS) generation rates.
    • OpenVINO™ Model Server simplifies model deployment and unifies REST API endpoints, reducing command complexity while providing standard v1/chat/completions support for easier integration with other serving frameworks.
    • OpenVINO™ Model Server further improves stability, performance, and accuracy for LLMs such as Qwen3.5/3.6 with linear attention and extends tool-parser support to MiniCPM5-1B and LFM2.5, improving accuracy and reliability.
Support Change and Deprecation Notices
  • Discontinued in 2026.0:

    • The deprecated openvino.runtime namespace has been removed. Please use the openvino namespace directly.
    • The deprecated openvino.Type.undefined has been removed. Please use openvino.Type.dynamic instead.
    • Support for Debian 10 has been discontinued due to the end of its standard support.
    • The PostponedConstant constructor signature has been updated for improved usability:
      • Old (removed): Callable[[Tensor], None]
      • New: Callable[[], Tensor]
    • The deprecated OpenVINO™ GenAI predefined generation configs were removed.
    • The deprecated OpenVINO GenAI support for whisper stateless decoder model has been removed. Please use a stateful model.
    • The deprecated OpenVINO GenAI StreamerBase put method, bool return type for callbacks, and ChunkStreamer class has been removed.
    • NNCF create_compressed_model() method is now deprecated and removed in 2026. Please use nncf.prune() method for unstructured pruning and nncf.quantize() for INT8 quantization.
    • NNCF optimization methods for TensorFlow models and TensorFlow backend in NNCF are deprecated and removed in 2026. It is recommended to use PyTorch analogous models for training-aware optimization methods and OpenVINO™ IR, PyTorch, and ONNX models for post-training optimization methods from NNCF.
    • The following experimental NNCF methods are deprecated and removed: NAS, Structural Pruning, AutoML, Knowledge Distillation, Mixed-Precision Quantization, Movement Sparsity.
    • CPU plugin now requires support for the AVX2 instruction set as a minimum system requirement. The SSE instruction set will no longer be supported.
    • OpenVINO™ migrated builds based on RHEL 8 to RHEL 9.
    • Dropped support for the TensorFlow Serving (TFS) API in OpenVINO Model Server 2026.3. KServe API is recommended for classic model deployments.
  • Deprecated and to be removed in the future:

    • auto shape and auto batch size (reshaping a model in runtime) will be removed in the future. OpenVINO’s dynamic shape models are recommended instead.
    • Starting with 2026.0 release major internal refactoring of the graph iteration mechanism has been implemented for improved performance and maintainability. The legacy path can be enabled by setting the ONNX_ITERATOR=0 environment variable. This legacy path is deprecated and will be removed in future releases.
    • OpenVINO Model Server:
      • The dedicated OpenVINO operator for Kubernetes and OpenShift is now deprecated in favor of the recommended KServe operator. The OpenVINO operator will remain functional in upcoming OpenVINO™ Model Server releases but will no longer be actively developed. Since KServe provides broader capabilities, no loss of functionality is expected. On the contrary, more functionalities will be accessible and migration between other serving solutions and OpenVINO Model Server will be much easier.
      • Support for Stateful models will be deprecated. These capabilities were originally introduced for Kaldi audio models which is no longer relevant. Current audio models support relies on the OpenAI API, and pipelines implemented via OpenVINO GenAI library.
      • Directed Acyclic Graph Scheduler will be deprecated in favor of pipelines managed by MediaPipe scheduler and will be removed in 2026.3. That approach gives more flexibility, includes wider range of calculators and has support for using processing accelerators.
      • OpenVINO™ GenAI:
        • start_chat() / finish_chat() APIs are deprecated and will be removed in a future major release. Pass a ChatHistory object directly to generate() instead.

You can find OpenVINO™ toolkit 2026.3 release here:

Acknowledgements

Thanks for contributions from the OpenVINO developer community: @abhijain1204fujitsu @Akihiro-Tabuchi @ankitm3k @aobolensk @AshutoshSinghIntel @daehyun99 @ggladilov @goyaladitya05 @Lagmator22 @pj3iL @pjordanandrsn @ssam18 @sshekhar563 @Sujanian1304 @tknose @vmadananth @wine99

Release documentation is available here: https://docs.openvino.ai/2026 Release Notes are available here: https://docs.openvino.ai/2026/about-openvino/release-notes-openvino.html

View originalPermalink

2026.2.1

Summary of improvements  

  • Fixed issue ID 187077: YOLO26 fails to compile on GPU
  • Resolved an issue in NPU plugin with shared L0 command queue causing violation of app/inference priorities

You can find OpenVINO™ toolkit 2026.2.1 release here:

View originalPermalink

2026.2.0
Summary of major features and improvements  
  • More GenAI coverage and framework integrations to minimize code changes
    • New models supported: Gemma 4 E2B and Gemma 4 E4B
      • Only on CPUs & GPUs: Qwen3-Coder-Next, Qwen3.5, Qwen3.6, Trinity-mini, LFM2-24B-A2B, LFM2-8B-A1B, LFM2.5-350M​
      • Only on CPUs: YOLO26
      • Only on GPUs: Gemma 4 31B and Gemma 4 26B-A4B
      • Extended to GPUs: GPT-OSS-120B
    • Scaled Dot-Product Attention (SDPA) path support added for LFM2 models
    • Support for Hugging Face Transformers v5.0, ensuring compatibility with the latest model architecture for enhanced interoperability.
  • Broader LLM model support and more model compression techniques
    • OpenVINO™ GenAI introduces extension support for loading custom extension libraries and registering unsupported operations via the extensions property. This gives developers the flexibility to run models with custom ops that OpenVINO doesn't support out of the box.
    • INT4 KV-cache compression is enabled for GPUs, with substantial memory reduction when KV cache size is significant, such as with large input prompts exceeding 32K tokens.
    • OpenVINO GenAI significantly reduces model loading times on GPU when using cache blobs — preventing bottlenecks for multi-stage AI pipelines, including agentic use cases that rely on multiple models.
    • Optimized IR read mode with independently managed constant buffers to reduce peak memory usage by avoiding unnecessary duplication of weight data unless required for correctness (Linux support added in this release).
    • Preview: Enhanced XAttention accuracy on CPUs and GPUs through by-channel INT8 KV-cache quantization (compared to by-token INT8 KV-cache), matching the default by-channel INT8 KV cache quantization when XAttention is not enabled.
  • More portability and performance to run AI at the edge, in the cloud, or locally.
    • OpenVINO™ GenAI extends its JavaScript API to include a Text-to-Speech pipeline and VLM samples for browser and Node.js developers.
    • Prompt Lookup Decoding extended to vision-language pipelines, delivering significantly faster token generation for multimodal workloads on Intel CPUs and GPUs.
    • OpenVINO™ GenAI now has a smaller runtime footprint after eliminating ICU DLL dependencies from tokenization, leading to reduced memory usage, faster startup, and easier deployment.
    • OpenVINO™ Model Server extends tool-calling support to Qwen 3.5 and 3.6 models to enable agentic AI use cases.
    • OpenVINO™ Model Server adds streaming transcription support for speech-to-text, reducing latency for real-time voice applications.
Support Change and Deprecation Notices
  • Discontinued in 2026.0:

    • The deprecated openvino.runtime namespace has been removed. Please use the openvino namespace directly.
    • The deprecated openvino.Type.undefined has been removed. Please use openvino.Type.dynamic instead.
    • The PostponedConstant constructor signature has been updated for improved usability:
      • Old (removed): Callable[[Tensor], None]
      • New: Callable[[], Tensor]
    • The deprecated OpenVINO™ GenAI predefined generation configs were removed.
    • The deprecated OpenVINO GenAI support for whisper stateless decoder model has been removed. Please use a stateful model.
    • The deprecated OpenVINO GenAI StreamerBase put method, bool return type for callbacks, and ChunkStreamer class has been removed.
    • NNCF create_compressed_model() method is now deprecated and removed in 2026. Please use nncf.prune() method for unstructured pruning and nncf.quantize() for INT8 quantization.
    • NNCF optimization methods for TensorFlow models and TensorFlow backend in NNCF are deprecated and removed in 2026. It is recommended to use PyTorch analogous models for training-aware optimization methods and OpenVINO™ IR, PyTorch, and ONNX models for post-training optimization methods from NNCF.
    • The following experimental NNCF methods are deprecated and removed: NAS, Structural Pruning, AutoML, Knowledge Distillation, Mixed-Precision Quantization, Movement Sparsity.
    • CPU plugin now requires support for the AVX2 instruction set as a minimum system requirement. The SSE instruction set will no longer be supported.
    • OpenVINO™ migrated builds based on RHEL 8 to RHEL 9.
    • manylinux2014 upgraded to manylinux_2_28. This aligns with modern toolchain requirements but also means that CentOS 7 will no longer be supported due to glibc incompatibility.
  • Deprecated and to be removed in the future:

    • Support for Ubuntu 20.04 has been discontinued due to the end of its standard support.
    • auto shape and auto batch size (reshaping a model in runtime) will be removed in the future. OpenVINO’s dynamic shape models are recommended instead.
    • With the release of Node.js v22, updated Node.js bindings are now available and compatible with the latest LTS version. These bindings do not support CentOS 7, as they rely on newer system libraries unavailable on legacy systems.
    • Starting with 2026.0 release major internal refactoring of the graph iteration mechanism has been implemented for improved performance and maintainability. The legacy path can be enabled by setting the ONNX_ITERATOR=0 environment variable. This legacy path is deprecated and will be removed in future releases.
    • OpenVINO Model Server:
      • The dedicated OpenVINO operator for Kubernetes and OpenShift is now deprecated in favor of the recommended KServe operator. The OpenVINO operator will remain functional in upcoming OpenVINO Model Server releases but will no longer be actively developed. Since KServe provides broader capabilities, no loss of functionality is expected. On the contrary, more functionalities will be accessible and migration between other serving solutions and OpenVINO Model Server will be much easier.
      • TensorFlow Serving (TFS) API support is planned for deprecation. With increasing adoption of the KServe API for classic models and the OpenAI API for generative workloads, usage of the TFS API has significantly declined. Dropping date is to be determined based on the feedback, with a tentative target of mid-2026.
      • Support for Stateful models will be deprecated. These capabilities were originally introduced for Kaldi audio models which is no longer relevant. Current audio models support relies on the OpenAI API, and pipelines implemented via OpenVINO GenAI library.
      • Directed Acyclic Graph Scheduler will be deprecated in favor of pipelines managed by MediaPipe scheduler and will be removed in 2026.3. That approach gives more flexibility, includes wider range of calculators and has support for using processing accelerators.
      • OpenVINO™ GenAI:
        • start_chat() / finish_chat() APIs are deprecated and will be removed in a future major release. Pass a ChatHistory object directly to generate() instead.

You can find OpenVINO™ toolkit 2026.2 release here:

Acknowledgements

Thanks for contributions from the OpenVINO developer community: @abbyssoul @aobolensk @AshutoshSinghIntel @dbermond @desertfury @FarseenSh @goyaladitya05 @KarSri7694 @Lagmator22 @mostafafaheem @om4rrr @pjordanandrsn @reeseliao @Sahilbhatane @ssam18 @sshekhar563 @Vishwa2684 @vmadananth

Release documentation is available here: https://docs.openvino.ai/2026 Release Notes are available here: https://docs.openvino.ai/2026/about-openvino/release-notes-openvino.html

View originalPermalink

2026.1.0
Summary of major features and improvements  
  • More GenAI coverage and framework integrations to minimize code changes
    • New models supported on CPUs & GPUs: Qwen3 VL
    • New models supported on CPUs: GPT-OSS 120B
    • Preview: Introducing the OpenVINO backend for llama.cpp, which enables optimized inference on Intel CPUs, GPUs, and NPUs. Validated on GGUF models such as Llama-3.2-1B-Instruct-GGUF, Phi-3-mini-4k-instruct-gguf, Qwen2.5-1.5B-Instruct-GGUF, and Mistral-7B-Instruct-v0.3.
    • New notebook: Unified VLM chatbot with video file support and interactive model switching across Qwen3-VL, Qwen2.5-VL, and LLaVa-NeXT-Video.
  • Broader LLM model support and more model compression techniques
    • OpenVINO™ GenAI adds TaylorSeer Lite caching for image and video generation, accelerating diffusion-transformer inference across Flux, SD3, and LTX-Video pipelines, aligned with Hugging Face Diffusers.
    • LTX-Video generation on GPU achieves end-to-end acceleration through fusion of RMSNorm and RoPE operators, significantly improving video generation performance.
    • OpenVINO™ GenAI adds dynamic LoRA support for Qwen3-VL and VL models with LLM, allowing developers to swap adapters at runtime for efficient serving of multiple model variants in production without reloading the base model.
    • Preview: The release-weights API for ov::Model enables memory reclamation during model compilation on NPUs, delivering dramatically lower peak memory consumption for edge and client deployments. Users must set this property in ov::Model, and it will be applied during compilation.
  • More portability and performance to run AI at the edge, in the cloud, or locally.
    • Introducing support for Intel® Core™ Series 3 processors (formerly codenamed Wildcat Lake) and Intel® Arc™ Pro B70 Graphics with 32GB memory for single-GPU inference on 20-30B parameter LLMs
    • Prompt Lookup Decoding extended to vision-language pipelines, delivering significantly faster token generation for multimodal workloads on Intel CPUs and GPUs.
    • OpenVINO™ GenAI now has a smaller runtime footprint after eliminating ICU DLL dependencies from tokenization, leading to reduced memory usage, faster startup, and easier deployment.
    • OpenVINO GenAI introduces WhisperPipeline for Node.js via its NPM package, delivering production-ready speech recognition with word-level audio-to-text transcription.
    • OpenVINO™ Model Server enhances support for Qwen3-MOE and GPT-OSS-20b models, delivering improved performance, accuracy, and robust concurrent request handling with continuous batching. These pre-optimized models are available on Hugging Face for easy deployment. Additionally, the Model Server introduces image inpainting and outpainting capabilities via the /image endpoint for AI image editing.
Support Change and Deprecation Notices
  • Discontinued in 2026.0:

    • The deprecated openvino.runtime namespace has been removed. Please use the openvino namespace directly.
    • The deprecated openvino.Type.undefined has been removed. Please use openvino.Type.dynamic instead.
    • The PostponedConstant constructor signature has been updated for improved usability:
      • Old (removed): Callable[[Tensor], None]
      • New: Callable[[], Tensor]
    • The deprecated OpenVINO™ GenAI predefined generation configs were removed.
    • The deprecated OpenVINO GenAI support for whisper stateless decoder model has been removed. Please use a stateful model.
    • The deprecated OpenVINO GenAI StreamerBase put method, bool return type for callbacks, and ChunkStreamer class has been removed.
    • NNCF create_compressed_model() method is now deprecated and removed in 2026. Please use nncf.prune() method for unstructured pruning and nncf.quantize() for INT8 quantization.
    • NNCF optimization methods for TensorFlow models and TensorFlow backend in NNCF are deprecated and removed in 2026. It is recommended to use PyTorch analogous models for training-aware optimization methods and OpenVINO™ IR, PyTorch, and ONNX models for post-training optimization methods from NNCF.
    • The following experimental NNCF methods are deprecated and removed: NAS, Structural Pruning, AutoML, Knowledge Distillation, Mixed-Precision Quantization, Movement Sparsity.
    • CPU plugin now requires support for the AVX2 instruction set as a minimum system requirement. The SSE instruction set will no longer be supported.
    • OpenVINO™ migrated builds based on RHEL 8 to RHEL 9.
    • manylinux2014 upgraded to manylinux_2_28. This aligns with modern toolchain requirements but also means that CentOS 7 will no longer be supported due to glibc incompatibility.
    • MacOS x86 is no longer supported.
    • APT & YUM Repositories Restructure: Starting with release 2025.1, users can switch to the new repository structure for APT and YUM, which no longer uses year-based subdirectories (like “2025”). The old (legacy) structure is unavailable starting 2026.0. Detailed instructions are available on the relevant documentation pages:
    • OpenCV binaries removed from Docker images.
  • Deprecated and to be removed in the future:

    • Support for Ubuntu 20.04 has been discontinued due to the end of its standard support.
    • auto shape and auto batch size (reshaping a model in runtime) will be removed in the future. OpenVINO’s dynamic shape models are recommended instead.
    • With the release of Node.js v22, updated Node.js bindings are now available and compatible with the latest LTS version. These bindings do not support CentOS 7, as they rely on newer system libraries unavailable on legacy systems.
    • Starting with 2026.0 release major internal refactoring of the graph iteration mechanism has been implemented for improved performance and maintainability. The legacy path can be enabled by setting the ONNX_ITERATOR=0 environment variable. This legacy path is deprecated and will be removed in future releases.
    • OpenVINO Model Server:
      • The dedicated OpenVINO operator for Kubernetes and OpenShift is now deprecated in favor of the recommended KServe operator. The OpenVINO operator will remain functional in upcoming OpenVINO Model Server releases but will no longer be actively developed. Since KServe provides broader capabilities, no loss of functionality is expected. On the contrary, more functionalities will be accessible and migration between other serving solutions and OpenVINO Model Server will be much easier.
      • TensorFlow Serving (TFS) API support is planned for deprecation. With increasing adoption of the KServe API for classic models and the OpenAI API for generative workloads, usage of the TFS API has significantly declined. Dropping date is to be determined based on the feedback, with a tentative target of mid-2026.
      • Support for Stateful models will be deprecated. These capabilities were originally introduced for Kaldi audio models which is no longer relevant. Current audio models support relies on the OpenAI API, and pipelines implemented via OpenVINO GenAI library.
      • Directed Acyclic Graph Scheduler will be deprecated in favor of pipelines managed by MediaPipe scheduler and will be removed in 2026.3. That approach gives more flexibility, includes wider range of calculators and has support for using processing accelerators.
      • OpenVINO™ GenAI:
        • start_chat() / finish_chat() APIs are deprecated and will be removed in a future major release. Pass a ChatHistory object directly to generate() instead.

You can find OpenVINO™ toolkit 2026.1 release here:

Acknowledgements

Thanks for contributions from the OpenVINO developer community: @andersendsa @arunthakur009 @ItzCobaltboy @KarSri7694 @Rahuldrabit @samthakur587 @satheeshbhukya @Shi-pra-19 @Sujanian1304

Release documentation is available here: https://docs.openvino.ai/2026 Release Notes are available here: https://docs.openvino.ai/2026/about-openvino/release-notes-openvino.html

View originalPermalink

2026.0.0
Summary of major features and improvements  
  • More GenAI coverage and framework integrations to minimize code changes
    • New models supported on CPUs & GPUs: GPT-OSS-20B, MiniCPM-V-4_5-8B, and MiniCPM-o-2.6​
    • New models supported on NPUs: MiniCPM-o-2.6. In addition, NPU support is now available on Qwen2.5-1.5B-Instruct, Qwen3-Embedding-0.6B, Qwen-2.5-coder-0.5B.​
    • OpenVINO™ GenAI now adds word-level timestamp functionality to the Whisper Pipeline on CPUs, GPUs, and NPUs, enabling more accurate transcriptions and subtitling in line with OpenAI and FasterWhisper implementations.​
    • Phi-3-mini FastDraft model is now available on Hugging Face to accelerate LLM inference on NPUs. FastDraft optimizes speculative decoding for LLMs.
  • Broader LLM model support and more model compression techniques
    • With the new int4 data-aware weight compression for 3D MatMuls, the Neural Network Compression Framework enables MoE LLMs to run with reduced memory, bandwidth, and improved accuracy compared to data-free schemes-delivering faster, more efficient deployment on resource-constrained devices.​
    • Preview: the Neural Network Compression Framework now supports per-layer and per-group Look-Up Tables (LUT) for FP8-4BLUT quantization. This enables fine-grained, codebook-based compression that reduces model size and bandwidth while improving inference speed and accuracy for LLMs and transformer workloads.
  • More portability and performance to run AI at the edge, in the cloud, or locally.
    • Preview: OpenVINO™ GenAI adds VLM pipeline support to enhance Agentic AI framework integration.​
    • OpenVINO GenAI now supports speculative decoding for NPUs, delivering improved performance and efficient text generation through a small draft model that is periodically validated by the full-size model.​
    • Preview: NPU compiler integration with the NPU plugin enables ahead-of-time and on-device compilation without relying on OEM driver updates. Developers can enable this feature for a single, ready-to-ship package that reduces integration friction and accelerates time-to-value.​
    • OpenVINO™ Model Server adds enhanced support for audio endpoint plus agentic continuous batching and concurrent runs for improved LLM performance in agentic workflows on Intel CPUs and GPUs.
Support Change and Deprecation Notices
  • Discontinued in 2026.0:

    • The deprecated openvino.runtime namespace has been removed. Please use the openvino namespace directly.
    • The deprecated openvino.Type.undefined has been removed. Please use openvino.Type.dynamic instead.
    • The PostponedConstant constructor signature has been updated for improved usability:
      • Old (removed): Callable[[Tensor], None]
      • New: Callable[[], Tensor]
    • The deprecated OpenVINO™ GenAI predefined generation configs were removed.
    • The deprecated OpenVINO GenAI support for whisper stateless decoder model has been removed. Please use a stateful model.
    • The deprecated OpenVINO GenAI StreamerBase put method, bool return type for callbacks, and ChunkStreamer class has been removed.
    • NNCF create_compressed_model() method is now deprecated and removed in 2026. Please use nncf.prune() method for unstructured pruning and nncf.quantize() for INT8 quantization.
    • NNCF optimization methods for TensorFlow models and TensorFlow backend in NNCF are deprecated and removed in 2026. It is recommended to use PyTorch analogous models for training-aware optimization methods and OpenVINO™ IR, PyTorch, and ONNX models for post-training optimization methods from NNCF.
    • The following experimental NNCF methods are deprecated and removed: NAS, Structural Pruning, AutoML, Knowledge Distillation, Mixed-Precision Quantization, Movement Sparsity.
    • CPU plugin now requires support for the AVX2 instruction set as a minimum system requirement. The SSE instruction set will no longer be supported.
    • OpenVINO™ migrated builds based on RHEL 8 to RHEL 9.
    • manylinux2014 upgraded to manylinux_2_28. This aligns with modern toolchain requirements but also means that CentOS 7 will no longer be supported due to glibc incompatibility.
    • MacOS x86 is no longer supported.
    • APT & YUM Repositories Restructure: Starting with release 2025.1, users can switch to the new repository structure for APT and YUM, which no longer uses year-based subdirectories (like “2025”). The old (legacy) structure is unavailable starting 2026.0. Detailed instructions are available on the relevant documentation pages:
    • OpenCV binaries removed from Docker images.
  • Deprecated and to be removed in the future:

    • Support for Ubuntu 20.04 has been discontinued due to the end of its standard support.
    • auto shape and auto batch size (reshaping a model in runtime) will be removed in the future. OpenVINO’s dynamic shape models are recommended instead.
    • With the release of Node.js v22, updated Node.js bindings are now available and compatible with the latest LTS version. These bindings do not support CentOS 7, as they rely on newer system libraries unavailable on legacy systems.
    • Starting with 2026.0 release major internal refactoring of the graph iteration mechanism has been implemented for improved performance and maintainability. The legacy path can be enabled by setting the ONNX_ITERATOR=0 environment variable. This legacy path is deprecated and will be removed in future releases.
    • OpenVINO Model Server:
      • The dedicated OpenVINO operator for Kubernetes and OpenShift is now deprecated in favor of the recommended KServe operator. The OpenVINO operator will remain functional in upcoming OpenVINO Model Server releases but will no longer be actively developed. Since KServe provides broader capabilities, no loss of functionality is expected. On the contrary, more functionalities will be accessible and migration between other serving solutions and OpenVINO Model Server will be much easier.
      • TensorFlow Serving (TFS) API support is planned for deprecation. With increasing adoption of the KServe API for classic models and the OpenAI API for generative workloads, usage of the TFS API has significantly declined. Dropping date is to be determined based on the feedback, with a tentative target of mid-2026.
      • Support for Stateful models will be deprecated. These capabilities were originally introduced for Kaldi audio models which is no longer relevant. Current audio models support relies on the OpenAI API, and pipelines implemented via OpenVINO GenAI library.
      • Directed Acyclic Graph Scheduler will be deprecated in favor of pipelines managed by MediaPipe scheduler and will be removed in 2026.3. That approach gives more flexibility, includes wider range of calculators and has support for using processing accelerators.

You can find OpenVINO™ toolkit 2026.0 release here:

Acknowledgements

Thanks for contributions from the OpenVINO developer community: @Sujanian1304 @Nishant-ZFYII @srinjoydutta03 @samthakur587 @IamRam3

Release documentation is available here: https://docs.openvino.ai/2026 Release Notes are available here: https://docs.openvino.ai/2026/about-openvino/release-notes-openvino.html

View originalPermalink

2026.0.1
  • Preview: NPU compiler integration with the NPU plugin enables ahead-of-time and on-device compilation without relying on OEM driver updates. This feature is enabled by default in this release package.
  • Known Issues
    • Component: optimum; ID: 179936
      Description: phi-4-multimodal instruct model isn’t functional when converted using optimum-cli as channel-wise one (with –group-size -1) with OpenVINO 2026.0. It’s recommended to use for the conversion OV 2025.4/OV 2025.4.1
    • Component: GenAI ; ID: 179754 Description: Text2VideoPipeline will crash with RuntimeError when calling generate(guidance_scale=1.0) after the model was reshaped or compiled with guidance_scale > 1.0; use guidance_scale >= 1.0001 as a workaround until the fix is available.
    • Component: Runtime; ID: 180693 Description: Qwen3-30B-A3B converted with newer transformers doesn’t work, recommend using transformers 4.55.4 for model conversion which was verified and worked.
    • Component: GenAI ; ID: 179973
      Description: Qwen2-vl, Qwen-2.5VL, Qwen3-VL dense models may not work through GenAI API with GPU, due internal issue on model transformation level
    • Component: Runtime ; ID: 180696 Description: 2nd (and further) latency degradation for Qwen3-MOE family, including lack of ability to fit a model on iGPU, due high memory consumption and potential graph corruption. Problem affects only IRs generated with 2026.0, former IRs generated with 2025.4 will work properly.
    • Component: Runtime ; ID: 179009 Description: Memory leak for static builds with HybridCRT enabled; impacts Windows only

You can find OpenVINO™ toolkit 2026.0.1 release here:

View originalPermalink