Xeon 6 6700/6500 series

Learn more

Intel® Xeon® 6 6500 & 6700 Series Processors

Essential and enhanced P-core performance for ANZ edge, mainstream servers, data centres and telco workloads.

Banner Icon
Xeon Banner Image
Xeon Banner Image

Intel® Xeon® 6 6500 & 6700 Series Processors

Purpose-built P-core performance for edge computing, mainstream enterprise workloads, data centres, and telco environments across Australia and New Zealand.

Intel Xeon 6 banner image

Intel® Xeon® 6 Hot Chips Press Briefing Visual

Intel® Xeon® 6 Platform Highlights

Granite Rapids: High‑Performance P‑Cores

  • Up to 128 Redwood Cove P‑cores per socket
  • 12-channel DDR5 and 136 PCIe 5.0 lanes
  • AMX & AVX-512 acceleration for HPC and AI

Perfect for latency-sensitive workloads like simulation and real-time inference.

6500P / 6700P: Balanced SKUs

  • Up to 86 P‑cores per socket
  • 4/8 socket support + 1S R1 for density
  • Enterprise-grade AI, cloud, DBs

Ideal for mid-tier servers with hybrid and cloud-native workloads.

Sierra Forest: Scalable E‑Cores

  • Up to 144 Crestmont E‑cores per socket
  • High-density, low power, multi-threading
  • Optimized for containers & edge compute

Best for microservices and cloud-native scaling.

Memory & Bandwidth Leadership

  • DDR5-8000 & MRDIMM modules
  • Up to 2.8× bandwidth vs Gen 4
  • NUMA optimized for AI + analytics

For AI, analytics, and large memory applications.

Platform Architecture

  • Tile-based compute + I/O (EMIB)
  • Shared sockets: LGA-4710 & LGA-7529
  • Unified platform for long-term lifecycle

Simplifies validation and upgrade paths.

Power & Sustainability

  • Up to 2.5× rack density improvement
  • Per-core telemetry & optimization
  • Designed for ESG & energy compliance

Supports sustainable, efficient operations.

An architecture designed for flexibility
An architecture designed for flexibility

An architecture designed for flexibility — Intel® Xeon® 6 at Hot Chips 2023

Meeting Diverse Compute Demands

Intel® Xeon® 6 processors with P-cores are engineered to meet the full spectrum of enterprise workloads—from AI and HPC to data services and microservices. With built-in AI acceleration, high memory bandwidth, and robust I/O, they deliver performance and efficiency across edge, cloud, and data centre environments.

  • AI Workloads: AMX acceleration in every core powers inference, training, and GenAI models without relying on GPUs
  • High-Performance Computing: Up to 128 P-cores per CPU and 504 MB L3 cache for compute-intensive simulations and modeling
  • Data Services: MRDIMM DDR5-8000 and CXL 2.0 support enable massive memory bandwidth for analytics and databases
  • Web & Microservices: High I/O throughput and flexible socket configurations support containerized and scale-out workloads

Which Intel Xeon 6 Processor Is Recommended for Your Workloads?

  • Xeon 6 6300 Series: Entry-level P-core CPUs for edge, storage, and control plane workloads
  • Xeon 6 6500 Series: Essential performance for mainstream servers, virtualization, and moderate AI tasks
  • Xeon 6 6700 Series: Enhanced performance for AI, HPC, and multi-socket data centre deployments
  • Xeon 6 6900 Series: Maximum core density and bandwidth for the most demanding AI and HPC environments

Why Intel Xeon 6 Processors with P-Cores Excel at AI

  • Up to 128 P-cores per CPU for parallel and sequential AI tasks
  • AMX acceleration with FP16, BF16, and INT8 support in every core
  • MRDIMM memory delivers up to 2.3× bandwidth vs previous gen
  • Up to 192 PCIe 5.0 lanes for GPU, SmartNIC, and NVMe connectivity
  • Advanced RAS features for uptime and compliance in AI systems

Intel Xeon 6 processors are trusted host CPUs for AI-accelerated systems, including NVIDIA DGX and Intel Gaudi platforms. They manage orchestration, preprocessing, and offloading—keeping GPUs fed and systems efficient.

AI that thrives on CPUs alone. Explore Intel Artificial Intelligence solutions at intel.com/ai.

Meeting Diverse Compute Demands

Intel® Xeon® 6 processors with P-cores are engineered to meet the full spectrum of enterprise workloads—from AI and HPC to data services and microservices. With built-in AI acceleration, high memory bandwidth, and robust I/O, they deliver performance and efficiency across edge, cloud, and data centre environments.

  • AI Workloads: AMX acceleration in every core powers inference, training, and GenAI models without relying on GPUs
  • High-Performance Computing: Up to 128 P-cores per CPU and 504 MB L3 cache for compute-intensive simulations and modeling
  • Data Services: MRDIMM DDR5-8000 and CXL 2.0 support enable massive memory bandwidth for analytics and databases
  • Web & Microservices: High I/O throughput and flexible socket configurations support containerized and scale-out workloads

Which Intel Xeon 6 Processor Is Recommended for Your Workloads?

  • Xeon 6 6300 Series: Entry-level P-core CPUs for edge, storage, and control plane workloads
  • Xeon 6 6500 Series: Essential performance for mainstream servers, virtualization, and moderate AI tasks
  • Xeon 6 6700 Series: Enhanced performance for AI, HPC, and multi-socket data centre deployments
  • Xeon 6 6900 Series: Maximum core density and bandwidth for the most demanding AI and HPC environments

Why Intel Xeon 6 Processors with P-Cores Excel at AI

  • Up to 128 P-cores per CPU for parallel and sequential AI tasks
  • AMX acceleration with FP16, BF16, and INT8 support in every core
  • MRDIMM memory delivers up to 2.3× bandwidth vs previous gen
  • Up to 192 PCIe 5.0 lanes for GPU, SmartNIC, and NVMe connectivity
  • Advanced RAS features for uptime and compliance in AI systems

Intel Xeon 6 processors are trusted host CPUs for AI-accelerated systems, including NVIDIA DGX and Intel Gaudi platforms. They manage orchestration, preprocessing, and offloading—keeping GPUs fed and systems efficient.

AI that thrives on CPUs alone. Explore Intel Artificial Intelligence solutions at intel.com/ai.

Take Advantage of Powerful AI Performance from a CPU

Intel® Xeon® 6 processors with P-cores are built to handle demanding AI workloads. With Intel® Advanced Matrix Extensions (Intel® AMX) supporting INT8, BF16, and FP16 data types, these CPUs accelerate inference, training, and midsize GenAI models while maintaining SLA targets.

As host CPUs in AI-accelerated systems, Xeon 6 P-cores deliver superior I/O, high single-threaded performance, and exceptional memory bandwidth and capacity.

Optimize Performance-per-Watt for General Purpose Workloads

With core counts ranging from 8 to 86 and up to 176 PCIe 5.0 lanes in dual-socket systems, Xeon 6 P-cores scale efficiently. Their nearly linear power-performance curve ensures optimal energy use even at high server utilization.

Performance-Core Workload Benefits

  • 40% average performance gain vs 5th Gen Xeon
  • Up to 3× better Llama2 performance with Intel® AMX
  • Up to 2× better HammerDB MySQL performance
  • 1.58× higher performance-per-watt at 40% server utilization
  • 1.45× higher database analytics performance
  • 1.5× better performance and efficiency for NGINX edge workloads
  • 2.5× better HPCG benchmark performance with MRDIMM

Generational & Competitive Gains

  • Up to 17× better ResNet-50 BSX performance vs 2nd Gen Xeon
  • 3.9× average general compute performance vs 2nd Gen Xeon
  • 17:1 server consolidation with Xeon 6900 vs 2nd Gen Xeon
  • 5:1 consolidation with Xeon 6700 vs 2nd Gen Xeon, with 40% TCO gain
  • 2× average AI performance vs AMD EPYC 9755
  • 1.52× higher HPCG performance vs AMD EPYC 9655

Take Advantage of Powerful AI Performance from a CPU

Intel® Xeon® 6 processors with P-cores are built to handle demanding AI workloads. With Intel® Advanced Matrix Extensions (Intel® AMX) supporting INT8, BF16, and FP16 data types, these CPUs accelerate inference, training, and midsize GenAI models while maintaining SLA targets.

As host CPUs in AI-accelerated systems, Xeon 6 P-cores deliver superior I/O, high single-threaded performance, and exceptional memory bandwidth and capacity.

Optimize Performance-per-Watt for General Purpose Workloads

With core counts ranging from 8 to 86 and up to 176 PCIe 5.0 lanes in dual-socket systems, Xeon 6 P-cores scale efficiently. Their nearly linear power-performance curve ensures optimal energy use even at high server utilization.

Performance-Core Workload Benefits

  • 40% average performance gain vs 5th Gen Xeon
  • Up to 3× better Llama2 performance with Intel® AMX
  • Up to 2× better HammerDB MySQL performance
  • 1.58× higher performance-per-watt at 40% server utilization
  • 1.45× higher database analytics performance
  • 1.5× better performance and efficiency for NGINX edge workloads
  • 2.5× better HPCG benchmark performance with MRDIMM

Generational & Competitive Gains

  • Up to 17× better ResNet-50 BSX performance vs 2nd Gen Xeon
  • 3.9× average general compute performance vs 2nd Gen Xeon
  • 17:1 server consolidation with Xeon 6900 vs 2nd Gen Xeon
  • 5:1 consolidation with Xeon 6700 vs 2nd Gen Xeon, with 40% TCO gain
  • 2× average AI performance vs AMD EPYC 9755
  • 1.52× higher HPCG performance vs AMD EPYC 9655

Intel® Xeon® 6 Product & Processor Briefs

Explore two complementary documents: the Product Brief offers a high-level overview of the Xeon® 6 family, while the Processor Brief dives deeper into architecture, performance tiers, and platform capabilities.

Xeon 6 Product Brief - First Page

Intel® Xeon® 6 Product & Processor Briefs

Explore two complementary documents: the Product Brief offers a high-level overview of the Xeon® 6 family, while the Processor Brief dives deeper into architecture, performance tiers, and platform capabilities.

Xeon 6 Product Brief - First Page

Clarity. Precision. Confidence.

DiGiCOR helps you cut through platform noise — whether you're navigating Intel® Xeon®, AMD EPYC™, NVIDIA GPUs, ARM-based solutions, or custom hybrid deployments. We guide you toward decisions you can trust and systems you can scale.

✔ Independent, vendor-neutral recommendations tailored to your workloads
✔ Workload-driven system design across AI, HPC, storage, and edge
✔ Compliant sourcing, assembly, and enterprise-grade deployment support
✔ Hardware you can rely on — built for growth, performance, and maintainability
Talk to a DiGiCOR Architect

Clarity. Precision. Confidence.

DiGiCOR helps you cut through platform noise — whether you're navigating Intel® Xeon®, AMD EPYC™, NVIDIA GPUs, ARM-based solutions, or custom hybrid deployments.

  • Independent, vendor-neutral recommendations tailored to your workloads
  • Workload-driven system design across AI, HPC, storage, and edge
  • Compliant sourcing, assembly, and enterprise-grade deployment support
  • Hardware you can rely on — built for growth, performance, and maintainability
Expert Infrastructure Guidance

Get expert infrastructure guidance — no vendor bias, just clarity.

Talk to a DiGiCOR Architect

Need Help? Fill in the form and We'll Get in Touch