Qwen3.8 125B (6B active) Flash-Next
on NVIDIA RTX 4090 · Strata · 65,536 ctx
Use cases
codingagenticlong-context
Summary
User reports Qwen3.8-Flash-Next at around 122 tok/s in regular chat on an RTX 4090.
Setup is Strata with IQ2_XS weights at 64K context, with system RAM at 45 of 48GB.
The user has not yet benchmarked coding agents or long-context workloads, and is weighing a $700 RAM upgrade to 128GB DDR4 against a $3,100 128GB Strix Halo machine for local multi-agent work.
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.
Models most run on the NVIDIA RTX 4090
- Qwen3.8 27B4-bit40.0 t/s · 3 runs
- 149.6 t/s · one run
- 129.0 t/s · one run
- Qwen2.5 7B4-bit158.9 t/s · one run
- Qwen3.6 27B4-bit25.0 t/s · one run
- Qwen3.8 27B4-bit35.9 t/s · one run
Bars are medians on one scale, 0 to 158.9 tokens/s. Each dot is one run and opens its report. Press ▶ to watch text arrive at a row's speed.
Everything reported on the NVIDIA RTX 4090 →Every Qwen3.8 report →Estimate Qwen3.8 on this card →