Rent a GPU

Can the NVIDIA RTX 3090 Ti run Qwen3.8 Flash Next?

RTX 3090 Ti × Qwen3.8 Flash Next8K context
NoDoes not fit in memory
Memory at Q4_K_M109.9 GB of 24 GB
Weights 109.1 GBKV cache 0.2 GBRuntime 0.6 GB
On the GPU21%of the Q4_K_M weights
In system RAM85.9 GBthe rest spills over
Offload speed~12tokens per second
  1. FP16
  2. Q8_0
  3. Q6_K
  4. Q5_K_M
  5. Q4_K_M
  6. Q3_K_M
  7. Q2_K

No. Even at Q4_K_M, Qwen3.8 Flash Next needs about 109.9 GB, but the NVIDIA RTX 3090 Ti has 24 GB usable. You could offload some layers to system RAM, but it will be very slow.

Try other context lengths

Run it in the cloud

Too big for the NVIDIA RTX 3090 Ti? Run Qwen3.8 Flash Next on a rented GPU instead.

Rent a cloud GPU by the hour. No upfront cost, stop any time. Not sure whether to rent or buy? Compare the cost.

Rent a GPU on RunPodRent a GPU on Vast.ai

Memory needed by quantization

At 8,192 tokens of context. The black line marks the 24 GB the NVIDIA RTX 3090 Ti can use.

QuantQualityMemoryFits?Speed
FP16Full quality360.8 GBToo big—
Q8_0Near-lossless192.1 GBToo big—
Q6_KExcellent148.4 GBToo big—
Q5_K_MVery good128.8 GBToo big—
Q4_K_MGood, most popular109.9 GBToo big—
Q3_K_MNoticeably degraded88.8 GBToo big—
Q2_KHeavily degraded68.3 GBToo big—

Measured speed from real RTX 3090 Ti owners

Tokens per second that people measured on their own NVIDIA RTX 3090 Ti, next to our estimate.

No one has measured Qwen3.8 Flash Next on the NVIDIA RTX 3090 Ti yet. The speeds above are estimates. Run the benchmark and yours becomes the first real number on this page.

Have an RTX 3090 Ti? Add your real speed

One command times Qwen3.8 Flash Next in your own Ollama, shows you exactly what it would send, and asks before sending anything. No sign-up, no personal data.

macOS and Linux

curl -fsSL https://gpufit.dev/bench.sh | sh

Windows (PowerShell)

irm https://gpufit.dev/bench.ps1 | iex

How the benchmark works

GPUs that run Qwen3.8 Flash Next comfortably

Check Mac Studio M3 Ultra (256 GB) prices

Models that run well on the NVIDIA RTX 3090 Ti

Frequently asked questions

Can NVIDIA RTX 3090 Ti run Qwen3.8 Flash Next?

No. Even at Q4_K_M, Qwen3.8 Flash Next needs about 109.9 GB, but the NVIDIA RTX 3090 Ti has 24 GB usable. You could offload some layers to system RAM, but it will be very slow.

How much VRAM does Qwen3.8 Flash Next need?

About 109.9 GB at Q4_K_M with 8,192 tokens of context, and about 360.8 GB at full FP16 precision.

How fast is Qwen3.8 Flash Next on the NVIDIA RTX 3090 Ti?

It does not fit in the NVIDIA RTX 3090 Ti's memory, so most layers would run from system RAM and generation would be very slow.