Consumer GPUIntel

Intel Arc A750 8GB

Curated Aggregate·2024-10-082 workloads · 2 records
MyAI Rating7.6tok/s · Phi-3 Mini

VRAM

8 GB

TDP

225 W

MSRP

$249

Perf/W

0.11 tok/s/W

Cost/1K tok

$0.11/M

Tested

2024-10-08

Quick answer

How fast is Intel Arc A750 8GB for local AI workloads?

Intel Arc A750 8GB hits 58.0 tok/s on Phi-3 Mini, its strongest benchmarked workload (batch 1, 4096-token context, 8GB VRAM, 225W TDP). It has 2 records across 2 workloads in our database, with a MyAI Rating of 7.6/10. Llama 3 70B Q4 needs ~40GB VRAM, so check the VRAM column before assuming feasibility.

Source: MyAIHardware benchmark database (bench-x3-a750-phi3)As of 2024-05-15

Overview

Budget Intel

Intel Arc A750 8GB — Alchemist architecture GPU. 8GB GDDR6 at 512 GB/s, 225W TDP. Intel's first-generation discrete GPU with Xe HPG architecture.

AI Usefulness

8GB limits to 7B Q4 at ~25-35 tok/s via IPEX-LLM. Budget experimentation card for Intel-curious builders. The software ecosystem is significantly more mature than at launch but still trails NVIDIA for AI workflows. Consider the Arc B580 for double the VRAM at similar pricing.

Verdict

Intel Arc A750 8GB with 8GB VRAM at 225W TDP, scored across 2 workloads with 2 benchmark records.

Best workload

Phi-3 Mini

58 tok/s

Quantization

Q4_K_M

4K context · batch 1

LLM Inference Performance

015304560Llama 38B Q4Phi-3 Mini

Benchmarks (2 workloads)

WorkloadScoreQuantContextσStatusTested
Llama 3 8B Q4

llm

24.0tok/sQ4_K_M2K, Curated Aggregate2024-10-08
Phi-3 Mini

llm

58.0tok/sQ4_K_M4K, Curated Aggregate2024-05-15

MyAI Score

Enthusiast
7.6/10

Intel Arc A750 8GB clears a 7.6/10 based on workload-normalized throughput, memory headroom, efficiency, value, trust, and coverage.

Throughput
248
Capability
109
Efficiency
37
Value
43
Trust
43
Coverage
15
Composite benchmark495 / 1000

Workload Fit

What models fit this 8GB card at different quantization levels.

Q4
Q8
FP16
7-8B
Excellent
Tight
Won't fit
13-14B
Tight
Won't fit
Won't fit
32B
Won't fit
Won't fit
Won't fit
70B
Won't fit
Won't fit
Won't fit
Top Benchmarks
Phi-3 Mini58 tok/s
Llama 3 8B Q424 tok/s

Source

Arc Alchemist 8GB IPEX-LLM.

View sourceHow we benchmark →

Public Trust Layer

Trust score

6/10

MyAI rating

7.6

Runs

1

Freshness

Stale

Source-linked row with explicit verification status.

Tested on 2024-05-15; 834 days old.

Open primary source

Best place to start

Where to buy

Search first

Retailer we'd check first

Amazon search

For mainstream AI hardware, Amazon usually updates street pricing, seller availability, and shipping speed faster than most comparison sites.

Search Amazon listings
  • +Use $249 as your price anchor unless the part is clearly supply-constrained or newly launched.
  • +Check the exact cooler, board partner, or memory configuration before buying. The silicon may match, but noise and thermals do not.
  • +If the Amazon price looks inflated, wait or compare against recent street pricing rather than paying a panic premium.

This is a fit-driven buy, not a hype buy. Price discipline matters more than launch excitement here.

Affiliate note: we do not have a device-level ASIN yet, so this opens tagged Amazon search results for the exact product name.