Qwen3.8 125B (6B active) Flash-Next
on Unknown GPU · Strata · 131,072 ctx
Summary
User reports Qwen3.8 Flash-Next running at about 50 t/s average decode and 1k prefill across a 131k context window on a 12GB VRAM / 32GB RAM configuration.
Setup is Strata with the default auto tuner and an IQ2_XS quant from Unsloth's new Dynamic v3.1 QFN quants.
The user says the quant retains much of its reasoning quality and that Strata makes Flash-Next runnable on modest hardware, with configurations targeting 8GB VRAM and 32GB RAM using the smallest quants.
Compare this run
Medians of plain runs: one device, one request, no speculative decoding, the whole model in the device's memory. The report above is one setup, so read it beside them and not as one of them.
Qwen3.8 125B · 6B active at 4-bit by hardware
- 129.2 t/s · one run
- 105.5 t/s · 6 runs
- 57.2 t/s · 2 runs
- 55.0 t/s · one run
- 36.7 t/s · one run
- NVIDIA DGX Spark128 GB29.8 t/s · 2 runs
Bars are medians on one scale, 0 to 232.0 tokens/s. Each dot is one run and opens its report. Press ▶ to watch text arrive at a row's speed.