Millions of Aussie Orders!

Millions of Aussie Orders!

The Ultimate Destination for Gamers & Enthusiasts

1HR Click & Collect

1HR Click & Collect

Available 7 Days!

Google Review Score

Google Review Score

Same Day Dispatch

Same Day Dispatch

On 1000's of Products!

Multi-Award Winners

Multi-Award Winners

Best Retailer & Builder

Suggestion box

Which new products would you like to see at PCCG?

Intel Arc Pro B70 GDDR6 32GB

33P01IB0BB | Intel
FREE SHIPPING
$1999
Check back later!
Buy now, pay later!
The Intel Arc Pro B70 GDDR6 32GB Workstation Graphics Card is an elite visual accelerator and AI processing powerhouse precisely optimised to deliver absolute next-gen performance across enterprise deep learning pipelines, heavy architectural CAD models, and advanced 3D rendering suites. At its centre is the revolutionary Xe2-HPG Battlemage architecture featuring 32 Xe-cores, 32 dedicated ray-tracing units, and 256 high velocity XMX AI engines that unleash a blistering 367 Peak INT8 TOPS for localised large language model (LLM) inference and real-time visualisation. To ensure data-intensive calculations and machine learning scripts remain flawlessly stabilised, this workstation graphics card boasts an immense 32GB pool of high-speed ECC GDDR6 VRAM mapped across a wide 256-bit interface, yielding a massive 608 GB/s memory bandwidth to easily load gargantuan datasets without bottlenecking. Operating on a modern PCIe 5.0 x16 native interface, the platform thrives under robust thermal management layouts engineered by board partners to eliminate thermal throttling during multi-day computational stress. Complete with certified independent software vendor (ISV) stability and future-ready, high-bandwidth connectivity via quad DisplayPort 2.1 outputs capable of driving multiple pristine 8K displays, this GPU serves as a durable, future proof asset for creators, developers, and data scientists requiring an ultra-low latency local workstation dashboard.

Features


Intel® Arc™ Pro B70 GPU

One Family. Endless Workstation Possibilities.

A high-performance workstation GPU engineered for demanding AI inference and complex professional workloads. Intel® Arc™ Pro B70 enables faster execution of advanced AI models, supports scalable multi-GPU deployments, and delivers the reliability of workstation-grade software and certifications, ideal for users pushing the limits of AI, rendering, and high-end creative applications.

Run larger models, process richer data, and get dependable workstation performance with Intel® Arc™ Pro B-Series. Combining large, dedicated memory, robust graphics power, and latest-generation XMX AI engines, Intel® Arc™ Pro B-Series accelerates rendering, video processing, and AI workloads. Scale seamlessly from everyday tasks to demanding AI projects with multi-GPU support. Work faster, iterate freely, and stay in control as your workloads grow.

  • 32GB Memory
  • 32 Xᵉ-cores
  • 367 pTOPS (GPU Peak TOPS (trillions of operations per second) represents the peak throughput when running XMX workloads with INT8 datatype and dense models. Performance may vary based on configuration.)
  • 608 GB/s Memory Bandwidth
  • 160W-290W Total Board Power (Intel® reference specification. Partner Designs may vary.)
  • Gen 5 PCIe (x16)

Powering the Next Generation of Workstations

Enabling new classes of professional workloads as AI, data, and advanced visualisation reshape workstations.

Fast Time to First Token for Quicker Answers

Experience up to 6.3x faster response time for multiple users/requests with Intel® Arc™ Pro B70. (vs. Nvidia RTX Pro 4000 (on Linux). Run on Ubuntu 25.04 | Docker Images: nvcr.io/nvidia/vllm:25.12-py3, intel/llm-scaler-vllm:1.3. See intel.com/performanceindex for details. Results may vary)

Peak GPU Performance for Professionals

Get up to 69% higher performance on a variety of workstation applications with Intel® Arc™ Pro B70. (vs. Intel® Arc™ Pro B60. See intel.com/performanceindex for details. Results may vary. SPEC®, and SPECviewperf® are registered trademarks of the Standard Performance Evaluation Corporation. See http://www.spec.org/spec/trademarks.html for more information.)

Higher Throughput for Multi-Agent Flows

Unlock up to 85% higher token throughput for multiple users/requests with Intel® Arc™ Pro B70. (vs. Nvidia RTX Pro 4000 (on Linux). Run on Ubuntu 25.04 | Docker Images: nvcr.io/nvidia/vllm:25.12-py3, intel/llm-scaler-vllm:1.3. See intel.com/performanceindex for details. Results may vary.)

Designed Around How Professionals Work

Intel Arc Pro B70 GDDR6 32GB Feature 1

Design & Engineering ISV-Certified Software

Expedite complicated design changes, speed up visualisation, and ensure precise viewport linework in tools like AutoCAD, Solidworks, Maya, 3ds Max, and Inventor.

Intel Arc Pro B70 GDDR6 32GB Feature 2

Developers

Build, test, and refine larger models locally with fewer constraints. Faster iteration and local inference help accelerate development while keeping sensitive datasets private.

Intel Arc Pro B70 GDDR6 32GB Feature 3

Animators and Artists

Bring ideas to life faster with responsive 2D and 3D workflows and AI-powered tools that simplify workflows, so creativity stays in motion.

Intel Arc Pro B70 GDDR6 32GB Feature 4

Motion Designers and Editors

Deliver high-quality visuals on tighter timelines with faster rendering, smoother effects, and AI-powered assistance for subtitles, object isolation, and audio cleanup.

Intel Arc Pro B70 GDDR6 32GB Feature 5

Architecture & Construction ISV-Certified Software

Accelerate the graphics performance of complex models to achieve seamless viewport interactions and responsiveness in architectural tools such as Revit, Enscape, MicroStation, and Rhino.

Intel Arc Pro B70 GDDR6 32GB Feature 6

Product Design ISV-Certified Software

Create detailed 3D models with responsive lighting, textures, and shadows essential for persuasive client presentations or design exploration, using tools like Catia, Fusion, or NX.

Powering Advanced AI and Pro Workflows

High-Capacity Memory for Larger AI Models

Up to 32 GB memory lets you run larger AI models and complex datasets.

Next-Generation Xe2 Architecture

Delivers 50% better performance per core compared to previous generation to accelerate content creation and AI applications.

XMX AI Engines for Accelerated Intelligence

Provides acceleration for cutting-edge AI applications and enhanced content creation.

Driver Support, Tuned to Your Workload

Optimised driver flexibility for professional development and workstation performance.

Linux Driver Support for Scalable Deployments

Robust Linux drivers enable scalable AI deployments across development environments.

Linux Support for Multi-GPU AI

Combine the power and memory of multiple GPUs to execute AI models in Linux that take 100+ GB of VRAM.

Dedicated Ray Tracing Units for Photorealism

Built-in ray tracing units accelerate photorealistic visuals in 3D rendering applications.

Xᵉ Media Engine for Faster Video Workflows

Xᵉ Media Engine provides comprehensive codec support for faster video encode, decode, and exports.

Certifications for Reliable, Trusted Performance

Built on decades of software collaboration, Intel® works closely with leading Independent Software Vendors (ISVs) to certify Intel® Arc™ Pro graphics hardware and drivers for professional applications. Through rigorous testing and validation, these certifications help ensure stable, reliable performance and optimised compatibility across demanding workflows, from design and engineering to media and content creation. (Subject to ISV certification completion and specific configuration requirements. Go to https://www.intel.com/content/www/us/en/support/topics/certified-graphics-hardware-and-drivers-by-isv.html for more details. Intel® Arc™ Pro graphics driver support for systems with Intel® Arc™ Pro GPUs may vary by provider. Check with system vendor for further details.)

Specifications


General Data

Model Number 33P01IB0BB
H.264 Hardware Encode/Decode Yes (Codec capabilities may vary by device and configuration. Contact your manufacturer to understand the enabled hardware acceleration and codec capabilities for individual devices.)
H.265 (HEVC) Hardware Encode/Decode Yes (Codec capabilities may vary by device and configuration. Contact your manufacturer to understand the enabled hardware acceleration and codec capabilities for individual devices.)
AV1 Encode/Decode Yes (Codec capabilities may vary by device and configuration. Contact your manufacturer to understand the enabled hardware acceleration and codec capabilities for individual devices.)

Essentials

Product Collection Intel® Arc™ Pro B-Series Graphics
Model Number B70
Code Name Products formerly Battlemage
Microarchitecture Xe2
Lithography Type TSMC N5
Marketing Status Launched
Launch Date Q1'26 (The date the product was first introduced.)
Warranty Period 3 yrs (Warranty document for this product is available on https://supporttickets.intel.com/warrantyinfo.)
Embedded Options Available No (The designation “Embedded Options Available” indicates that a SKU is suitable for edge or embedded applications. For more information about edge or embedded use cases for this product, visit the Intel Resource and Documentation Center at (https://rdc.intel.com) or contact your Intel representative. For details on how a specific Intel product is used in an edge or embedded device, please contact the device manufacturer.)

GPU

Xe-cores 32
Render Slices 8
Ray Tracing Units 32
Intel® Xe Matrix Extensions (Intel® XMX) Engines 256
Xe Vector Engines 256
Graphics Max Dynamic Clock 2800 MHz
Graphics Clock 2280 MHz (The graphics clock represents the average clock an end user may see in a typical gaming workload.)
GPU TFLOPS (FP32) 22.94 (This value represents the highest-measured rate of 32-bit floating-point operations (FP32) on this product in trillions of floating-point operations per second (TFLOPS).
GPU Peak TOPS (Int8) 367 (GPU Peak TOPS (trillions of operations per second) represents the peak throughput when running XMX workloads with INT8 datatype and dense models. Performance may vary based on configuration.)
TBP 230 W (Total Board Power (TBP) represents the total power draw of a graphics card or other add-in card in watts, when it is operating under a typical load such as a gaming workload. The TBP value corresponds to Intel’s reference design. Intel’s partners may choose to productise Intel-based solutions with higher TBP values.)
TBP - LP 160-290 (Total Board Power (TBP) – Low Power (LP) represents the total power draw of a graphics card or other add-in card in watts, when it is operating under a typical workload. The TBP range corresponds to Intel’s reference design in a low-power configuration. Intel’s partners may choose to productise Intel-based solutions with different TBP values.)
PCI Express Configurations PCI Express 5.0 x16 (This feature may not be available on all computing systems. Please check with the system vendor to determine if your system delivers this feature, or reference the system specifications (motherboard, processor, chipset, power supply, HDD, graphics controller, memory, BIOS, drivers, virtual machine monitor-VMM, platform software, and/or operating system) for feature compatibility. Functionality, performance, and other benefits of this feature may vary depending on system configuration.) (PCI Express (PCIe) Configurations describe the available PCIe lane configurations that can be used to link to PCIe devices.)
Device ID 0xE223

Memory

Memory 32 GB GDDR6
Graphics Memory Interface 256 bit
Graphics Memory Bandwidth 608 GB/s
ECC Memory Supported Yes (This feature may not be available on all computing systems. Please check with the system vendor to determine if your system delivers this feature, or reference the system specifications (motherboard, processor, chipset, power supply, HDD, graphics controller, memory, BIOS, drivers, virtual machine monitor-VMM, platform software, and/or operating system) for feature compatibility. Functionality, performance, and other benefits of this feature may vary depending on system configuration.) (ECC Memory Supported indicates processor support for Error-Correcting Code memory. ECC memory is a type of system memory that can detect and correct common kinds of internal data corruption. Note that ECC memory support requires both processor and chipset support.)

I/O

Number of Displays Supported 4 (This feature may not be available on all computing systems. Please check with the system vendor to determine if your system delivers this feature, or reference the system specifications (motherboard, processor, chipset, power supply, HDD, graphics controller, memory, BIOS, drivers, virtual machine monitor-VMM, platform software, and/or operating system) for feature compatibility. Functionality, performance, and other benefits of this feature may vary depending on system configuration.)
Graphics Output HDMI* 2.1 (This feature may not be available on all computing systems. Please check with the system vendor to determine if your system delivers this feature, or reference the system specifications (motherboard, processor, chipset, power supply, HDD, graphics controller, memory, BIOS, drivers, virtual machine monitor-VMM, platform software, and/or operating system) for feature compatibility. Functionality, performance, and other benefits of this feature may vary depending on system configuration.)
HDMI Variable Refresh Rate (VRR) Yes
Max Resolution (HDMI) 7680 x 4320@120Hz (This feature may not be available on all computing systems. Please check with the system vendor to determine if your system delivers this feature, or reference the system specifications (motherboard, processor, chipset, power supply, HDD, graphics controller, memory, BIOS, drivers, virtual machine monitor-VMM, platform software, and/or operating system) for feature compatibility. Functionality, performance, and other benefits of this feature may vary depending on system configuration.) (Max Resolution (HDMI) is the maximum resolution supported by the processor via the HDMI interface (24bits per pixel & 60Hz). System or device display resolution is dependent on multiple system design factors; actual resolution may be lower on your system.)
Max Resolution (DP) 7680 x 4320@120Hz (This feature may not be available on all computing systems. Please check with the system vendor to determine if your system delivers this feature, or reference the system specifications (motherboard, processor, chipset, power supply, HDD, graphics controller, memory, BIOS, drivers, virtual machine monitor-VMM, platform software, and/or operating system) for feature compatibility. Functionality, performance, and other benefits of this feature may vary depending on system configuration.) (Max Resolution (DP) is the maximum resolution supported by the processor via the DP interface (24bits per pixel & 60Hz). System or device display resolution is dependent on multiple system design factors; actual resolution may be lower on your system.)

Supported Technologies

Ray Tracing Yes
oneAPI Support Yes (oneAPI is an open, cross-industry, standards-based, unified, multiarchitecture, multi-vendor programming model that delivers a common developer experience across accelerator architectures – for faster application performance, more productivity, and greater innovation. The oneAPI initiative encourages collaboration on the oneAPI specification and compatible oneAPI implementations across the ecosystem.)
OpenVINO™ Support Yes (OpenVINO stands for “Open Visual Inference and Neural network Optimisation” It is a free toolkit that provides developers with improved neural network performance on a variety of Intel devices and helps them further unlock cost-effective, real-time vision applications. The toolkit enables deep learning inference and easy heterogeneous execution across multiple Intel platforms (CPU, Intel® Processor Graphics, Movidius VPU, FPGA and more) — providing implementations across cloud architectures to edge devices. Learn more about the Intel® Distribution of OpenVINO™ Toolkit.)
DirectX* Support DirectX 12 Ultimate (DirectX* Support indicates support for a specific version of Microsoft’s collection of APIs (Application Programming Interfaces) for handling multimedia compute tasks.)
Vulkan* Support 1.3
OpenGL* Support Up to 4.6 (OpenGL (Open Graphics Library) is a cross-language, multi-platform API (Application Programming Interface) for rendering 2D and 3D vector graphics.)
OpenCL* Support 3.0 (OpenCL (Open Computing Language) is a multi-platform API (Application Programming Interface) for heterogeneous parallel programming.)
Media Profiles HDR10, HDR10+ GAMING, Dolby Vision
Multi-Format Codec Engines 2 (Multi-Format Codec Engines provide hardware encoding and decoding for amazing video playback, content creation, and streaming uses.)
VESA Adaptive Sync Yes
Intel® Extension for PyTorch* (IPEX) Support Yes (Intel® Extension for PyTorch* (IPEX) Support", boolean, restriction 0, bottom of Supported Technologies. Tooltip: "The Intel® Extension for PyTorch* for GPU extends PyTorch with up-to-date features and optimisations for an extra performance boost on Intel Graphics cards. Intel releases its newest optimisations and features in Intel® Extension for PyTorch* before upstreaming them into open source PyTorch. Read more about Intel® Extension for PyTorch*.)
Intel® Xe Super Sampling (XeSS) Support Yes (Intel® Xe Super Sampling (XeSS) is AI-enhanced upscaling, enabling more performance with high image fidelity that can leverage the DP4a and XMX hardware capabilities of Xe GPUs. XeSS is implemented using open standards like Shader Model 6.4 to ensure wide availability on many games and across a broad set of shipping hardware. Read more about Intel® Xe Super Sampling.)

Intel Reference Card Attributes

Dimensions (Length x Width) 10.5" x 3.9"
Slots 2
Weight 1020 G
Power Connectors 1 ATX 2X4 (8Pin)
Backed by a 3 year Intel wararnty.

Customer reviews

Customer questions