Can I run it?
Can I run Qwen3.8 Flash Next on RTX 6000 Ada?
Partial offload
Partial offload on your RTX 6000 Ada: UD-IQ1_M at 4K context, 30 of 48 layers on the GPU and 25.4 GB in system RAM. Expect about 3.8–5.2 tok/s.
Best quantUD-IQ1_M
Working context4K
Max that fits4K
Speed3.8–5.2 tok/sEstimate from memory bandwidth, not a benchmark. Method.
Weights69.4 GB
KV cache0.15 GB
Total70.6 GB
Usable45.1 GB
Qwen3.8 Flash Next details · RTX 6000 Ada page · How this estimate is made
Run commands
llama.cpp
llama-server -hf unsloth/Qwen3.8-Flash-Next-GGUF:UD-IQ1_M -c 4096 -ngl 30
Ollama
ollama run qwen3.8-flash-next
LM Studio
lms get unsloth/Qwen3.8-Flash-Next-GGUF # Then load quant UD-IQ1_M
Assumptions
- Estimate from memory bandwidth, not a benchmark. Decode speed uses 55–75% of peak bandwidth.
- KV cache is FP16. Engine overhead assumes llama.cpp (0.3 GB + 1% of the weights). Ollama is closer to 0.9 GB + 3%, LM Studio closer to 1 GB + 3%.
- UD-IQ1_M size is the real GGUF file size from Hugging Face.
- 0.5 GB is reserved because this GPU also drives a display, then 5% of the rest is left free.
- Speed is quoted at 4K context, not at the maximum context that still fits.
- This quant is IQ1 or IQ2. It can load, but quality is heavily degraded, so it is left off best-fit lists.
- About 30 of 48 layers stay on the GPU. 25.4 GB is read from system RAM.
Models that fit this machine more comfortably
- Mistral Medium 3.5 128B — Heavily degraded, UD-IQ2_XXS, 14–19 tok/s (estimate)
- Qwen3.5 122B A10B MTP — Tight fit, UD-Q2_K_XL, 30–41 tok/s (estimate)
- Mistral Small 4 119B 2603 — Tight fit, UD-IQ3_S, 29–39 tok/s (estimate)
Updated 2026-10-24. Profile on this page is the device default plus 32 GB of DDR5-5600 system RAM.