How much VRAM does the Intel Arc A310 4GB have?

The Intel Arc A310 4GB has 4 GB of GDDR6 with 124 GB/s memory bandwidth.

What is the Intel Arc A310 4GB best for?

With 4 GB of VRAM, the Intel Arc A310 4GB is best for running compact models (1B–8B) at low quantization, suitable for edge inference, prototyping, and lightweight tasks.

What LLMs can the Intel Arc A310 4GB run locally?

The Intel Arc A310 4GB can run 13 of the 70 open-weight models tracked by CanItRun natively in VRAM at 8k context. Top options include: Llama 3.2 3B Instruct at Q5_K_M, Llama 3.2 1B Instruct at BF16, Qwen 2.5 7B Instruct at Q2_K.

Can the Intel Arc A310 4GB run Llama 3.3 70B Instruct?

The Intel Arc A310 4GB can run Llama 3.3 70B Instruct with CPU offload at Q2_K quantization, but inference will be slower than native VRAM execution.

Can the Intel Arc A310 4GB run Qwen 3.6 27B?

The Intel Arc A310 4GB can run Qwen 3.6 27B with CPU offload at Q6_K quantization, but inference will be slower than native VRAM execution.

Can the Intel Arc A310 4GB run Llama 3.1 8B Instruct?

The Intel Arc A310 4GB can run Llama 3.1 8B Instruct with CPU offload at BF16 quantization, but inference will be slower than native VRAM execution.

Intel Arc A310 4GB

The Intel Arc A310 4GB has 4 GB VRAM and 124 GB/s memory bandwidth. It can run 13 of our 70 tracked models natively in VRAM at 8k context.

Intel Arc A310 4GB: 2022 Xe-HPG Alchemist with 4GB GDDR6 at 124 GB/s — Intel's entry-level Arc card.

3B models at Q4 with tight fit; 4GB is very limiting for LLM inference. ~1-3 t/s for 3B via Vulkan.

Vulkan via llama.cpp works cross-platform. SYCL backend requires oneAPI toolkit. Ollama support limited.

Vendor	Intel
Architecture	Xe-HPG (Alchemist)
VRAM	4 GB
Memory type	GDDR6
Memory bandwidth	124 GB/s
Compute backend	VULKAN
Tier	Consumer
Released	2022
Models (native)	13 / 70
Models (offload)	34 / 70

Software: Vulkan backend works in llama.cpp; SYCL backend available but requires oneAPI toolkit. Ollama support is limited.

Models this GPU runs natively in VRAM (13)

Models that fit with CPU offload (34)

These use system RAM for layers that don't fit in VRAM — expect much slower inference.

Too large for this GPU (23)

Frequently asked questions

How much VRAM does the Intel Arc A310 4GB have?: The Intel Arc A310 4GB has 4 GB of GDDR6 with 124 GB/s memory bandwidth.
What is the Intel Arc A310 4GB best for?: With 4 GB of VRAM, the Intel Arc A310 4GB is best for running compact models (1B–8B) at low quantization, suitable for edge inference, prototyping, and lightweight tasks.
What LLMs can the Intel Arc A310 4GB run locally?: The Intel Arc A310 4GB can run 13 of the 70 open-weight models tracked by CanItRun natively in VRAM at 8k context. Top options include: Llama 3.2 3B Instruct at Q5_K_M, Llama 3.2 1B Instruct at BF16, Qwen 2.5 7B Instruct at Q2_K.
Can the Intel Arc A310 4GB run Llama 3.3 70B Instruct?: The Intel Arc A310 4GB can run Llama 3.3 70B Instruct with CPU offload at Q2_K quantization, but inference will be slower than native VRAM execution.
Can the Intel Arc A310 4GB run Qwen 3.6 27B?: The Intel Arc A310 4GB can run Qwen 3.6 27B with CPU offload at Q6_K quantization, but inference will be slower than native VRAM execution.
Can the Intel Arc A310 4GB run Llama 3.1 8B Instruct?: The Intel Arc A310 4GB can run Llama 3.1 8B Instruct with CPU offload at BF16 quantization, but inference will be slower than native VRAM execution.