GPUs / NVIDIA

RTX 4060

Memory
GDDR6
Bus
128-bit
Published peak
272 GB/s
Capacity
8 GB

Measured

Read
250.5GB/s
Write
244.6GB/s
Copy
244.1GB/s
Random
39.1GB/s
Latency
235ns
Sustain
246.6GB/s

Medians of 5 verified runs from 1 operator. Read reaches 92 % of the published peak.

68.85 verified runs, median 69.269.5

Local AI speed

ModelParameters4-bit, tokens/s8-bit, tokens/s
Llama 3.1 8B8.03B≤ 62Does not fit
Qwen 2.5 14B14.7B≤ 34Does not fit
gpt-oss-20b21B, 3.6B activeDoes not fitDoes not fit
Mistral Small 24B24BDoes not fitDoes 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 8 GB configuration; the KV cache needs room on top.

Top operators

#OperatorScoreReadDate
1OP-62274469.5264.3 GB/s2026-09-27

Full RTX 4060 leaderboard