2x RTX 3060 12 GB

24 GB · 12 GB each · dual-card layer split · $614 Card · $1,814 Full rig

Quick decision verdict

2x RTX 3060 12 GB outputs 41.3 tok/s at $0.101/M tokens (rank #17 of 28 measured rigs on unit cost), breaking even in 738 days at full load.

  • Generation speed rank #22 of 29 · System power rank #20 (490 W).
  • For a lower upfront build cost, compare against AMD Radeon RX 6900 XT 16GB ($1,669)
  • If you need faster interactive token speed, compare against RTX PRO 6000 Blackwell (171.1 tok/s)

Can I Run It? → · Best → · Method →

How you'll use it

Fits?Source
2x RTX 3060 12 GB
12 GB each · dual-card layer split
24 GB1,239.841.3Roomy490 W$1,81427.5$0.100$1.76$4.23736 daysMeasured

Selected hardware

2x RTX 3060 12 GB

24 GB · 12 GB each · dual-card layer split · Layer split across two cards

Read tok/s
1,239.8
Write tok/s
41.3
Power
490 W
Output/day
27.5
Cost/M tok
$0.100
Power/day
$1.76
Saved/day
$4.23
Payback
736 days

Details · See source

FAQ

How fast does 2x RTX 3060 12 GB run local LLMs?
2x RTX 3060 12 GB (24 GB) measures 41.3 tokens/s generation and 1,240 tokens/s prompt reading on Qwen3.8-Flash-Next (snapshot 2026-10-07).
Is 2x RTX 3060 12 GB worth buying over cloud LLM APIs?
At $0.101/M tokens local cost, 2x RTX 3060 12 GB breaks even in 738 days at 24h/day utilization, or 4,418 days at 4h/day — making daily utilization the main deciding factor.