ModelDock
1
Mac Studio M5 Ultra 256 GBEstimate
~1,012.2
~70.4
256 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 2026-08-28
Hardware × model dataset
Published Qwen3.8 Flash Next hardware performance
2
RTX 5090Partial estimate
~1,301
65
32 GB
3
Mac Studio M5 Ultra 96 GBEstimate
~828.6
~57.6
96 GB
517.9
36
64 GB
05
Mac Studio M5 Max 128 GBEstimate
~517.9
~36
128 GB
06
Mac Studio M5 Max 64 GBEstimate
~517.9
~36
64 GB
07
RTX 4090Partial estimate
~721
36
24 GB
540.1
21.1
128 GB
09
Mac mini M5 Pro 64 GBEstimate
~259
~18
64 GB
10
RTX 3090Partial estimate
~116
5.8
24 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.