01
GPU card
6,256.6
173.8
32 GB
ModelDock / model benchmark
Qwen3.8 Flash Next local hardware benchmarks
Compare Qwen3.8 Flash Next reading / prefill and single-stream writing / decode performance across source-backed hardware and clearly labelled projections. Model profile: 125B main + 51B N-gram PLE · 6B active.
ModelQwen3.8 Flash Next125B main + 51B N-gram PLE · 6B active · Dataset as of 2026-10-15
Hardware × model dataset
Published Qwen3.8 Flash Next hardware performance
02
GPU card
8,987.9
171.1
96 GB
03
GPU card
3,441.5
155
256 GB
04
GPU card
4,083
139.6
96 GB
05
GPU card
2,137.6
111.3
48 GB
06
GPU card
4,000
109.3
24 GB
07
GPU card
3,030
96
32 GB
08
RTX 3090 TiEstimate
GPU card
~2,255.1
~93.1
24 GB
09
GPU card
2,094
86.5
24 GB
10
GPU card
2,197.4
80.8
32 GB
11
GPU card
3,591.4
80.1
16 GB
12
GPU card
4,244.5
79.2
64 GB
13
GPU card
2,355.3
78.8
64 GB
14
GPU card
1,245.5
77.6
128 GB
15
GPU card
1,700
72
16 GB
16
GPU card
1,529.2
71.2
24 GB
17
GPU card
2,498.5
71.1
16 GB
18
GPU card
1,602
67.6
32 GB
19
GPU card
1,713.4
66.1
48 GB
20
GPU card
1,996.3
66
16 GB
21
GPU card
2,146.5
65.2
128 GB
22
GPU card
772.3
63
32 GB
23
GPU card
1,660
54.9
16 GB
24
GPU card
782.8
51.4
32 GB
25
GPU card
2,228.5
48.7
128 GB
26
GPU card
1,443.2
48.3
12 GB
27
GPU card
466
47.6
16 GB
28
GPU card
1,242.7
47.5
128 GB
29
GPU card
1,239.8
41.3
24 GB
30
GPU card
1,376
36.9
12 GB
31
GPU card
517.9
36
64 GB
32
Mac Studio M5 Max 64 GBEstimate
GPU card
~517.9
~36
64 GB
33
GPU card
1,034
35.5
24 GB
34
GPU card
446.8
32.5
128 GB
35
GPU card
820
20.2
12 GB
36
GPU card
430
19.2
64 GB
37
GPU card
479.6
18.1
512 GB
Reading / prefill
Prefill · tok/s
Writing / decode
Decode · single stream · tok/s
Evidence
Rows use the same hardware × model projection as the live Benchmark and Ranking interfaces. Estimates remain visibly marked.