GPUs / Intel

Arc 140T

Memory
LPDDR5X / DDR5
Bus
128-bit
Published peak
120 to 136.5 GB/s
Capacity
from 16 GB

Measured

Read
37.6GB/s
Write
31.2GB/s
Copy
36.0GB/s
Random
6.91GB/s
Latency
748ns
Sustain
34.7GB/s

Medians of 1 verified run from 1 operator.

Local AI speed

ModelParameters4-bit, tokens/s8-bit, tokens/s
Llama 3.1 8B8.03B≤ 9.4≤ 4.7
Qwen 2.5 14B14.7B≤ 5.1≤ 2.6
gpt-oss-20b21B, 3.6B active≤ 21Does not fit
Mistral Small 24B24B≤ 3.1Does not fit
Qwen 2.5 32B32.5BDoes not fitDoes not fit
Llama 3.3 70B70.6BDoes not fitDoes not fit
gpt-oss-120b117B, 5.1B activeDoes not fitDoes not fit
Upper bounds: median read ÷ bytes read per token. Mixture-of-experts models read only their active experts. Fit checks weights against the smallest 16 GB configuration; the KV cache needs room on top.

Top operators

#OperatorScoreReadDate
1OP-B049F528.537.6 GB/s2026-09-28

Full Arc 140T leaderboard