[ENG] ASUS Announces UGen Intel Arc Pro B70 32GB AI Accelerator Cards

Designed for the rise of local agentic AI

Modern AI workloads can be memory-hungry. Long-context reasoning dramatically inflates KV-cache usage, and multi-agent deployments run several models simultaneously, multiplying VRAM requirements. Once VRAM runs short, data offloads to host memory, and inference performance can drop by up to 90%. This card’s 32GB of ECC-protected GDDR6 memory can be enough to keep demanding models entirely local. A single ASUS UGen B70 supports most models under 27 billion parameters and enables a context window of up to 128K tokens with Llama 3.1 8B, scaling to 64GB in dual-card configurations and 128GB in quad configurations to meet the AI deployment needs of SMBs and large enterprises.

A mature software ecosystem

The ASUS UGen B70 is supported by a broad and rapidly maturing AI software stack, including Intel LLM Scaler and vLLM XPU, llama.cpp (SYCL and Vulkan), Ollama, Intel OpenVINO and oneAPI, PyTorch XPU, and ComfyUI for image generation. Verified models span Llama 3.1/3.2 (8B–70B), Qwen 3.x, DeepSeek-R1 Distill (7B–70B), Mistral and Mixtral, OpenAI gpt-oss (20B and 120B), Whisper-large-v3, and InternVL3.5 — with dual-card configurations proven to run 80B-class models at 43.4 tokens per second. LLM Scaler’s built-in INT4/FP8 online quantization removes the dependency on external CUDA-based quantization tools, and currently known ecosystem limitations do not affect near-term commercial deployment.

Designed for AI adoption at every scale

The ASUS UGen B70 addresses three key user segments. For prosumers and independent AI developers, a single 32GB card breaks through the resource ceiling that mainstream 16GB cards impose on larger models. For small businesses and creative studios, dual-card 64GB configurations support larger deployments while keeping sensitive data in-house. For enterprise departmental AI teams, multi-card 128GB systems dramatically reduce long-term costs compared with cloud APIs while satisfying data-compliance requirements in fields such as law, medicine, government, and finance.

Across every tier, this scalability is what makes the ASUS UGen B70 the building block of an agent computer – a system designed from the ground up to run AI agents locally.

Learn more about Agent Computer – ASUS's vision for powering every AI choice, across cloud, local, and hybrid AI deployments: https://www.asus.com/microsite/agent-computer/

Availability & Pricing

The ASUS UGen Intel Arc Pro B70 32GB is available in Germany, Austria and Switzerland at a recommended retail price of 1,799.- EUR / CHF incl. VAT from specialist retailers and online shops now.