Qwen3.8-27B · Q2_K
This page aggregates 1 real-world runs of Qwen3.8-27B (Q2_K) on NVIDIA RTX 5090 with vLLM, contributed by 1 independent source platforms; metrics are averages of published measurements.
248.25 tok/s
Decode
Decode speed
794.5 tok/s
Prefill
Prefill speed
1.42 s
TTFT
Time to first token
30.4 GB
VRAM
VRAM usage
1
Measured runs
1
Independent sources
Other
Source platforms
Today
Last verified
Performance
- vLLM · Q2_K (current)Decode 248.25 · Prefill 794.5 ·
- Ollama · Q2_K Decode 231.24 · Prefill 1001.06 ·
- llama.cpp · IQ4_XS Decode 95.7 · Prefill 717.8 ·
- llama.cpp · Q5_K_S Decode 76.8 · Prefill 576 ·
- llama.cpp · IQ3_XXS Decode 147.1 · Prefill 1103.3 ·
- llama.cpp · IQ3_S Decode 128.2 · Prefill 961.5 ·
- llama.cpp · Q6_K Decode 59 · Prefill 442.5 ·
- vLLM · Q4_K_M Decode 92.9 · Prefill 696.8 ·
Core figures
- Decode (avg)
- 248.25 tok/s
- Prefill (avg)
- 794.5 tok/s
- TTFT (avg)
- 1.42 s
- VRAM (avg)
- 30.4 GB
- MTP acceptance rate
- —
- TTFB
- 1.99 GB
- Power draw
- 436.7 W
Configuration
Member-level fields are taken from the most recent run
- Model
- Qwen3.8-27B
- Quantization
- Q2_K
- Framework
- vLLM
- Version
- v0.24
- Context length
- 8192 tokens
- Batch size
- 32
- GPU layers
- 999
- Flash Attention
- On
Hardware
Nominal and measured figures are shown side by side; whether it runs is the reader's call
- GPU
- NVIDIA RTX 5090
- Nominal VRAM
- 32 GB
- Measured VRAM (avg)
- 30.4 GB
- OS
- Ubuntu 22.04
- Driver
- 570.86
- CUDA
- 12.8
- Power draw
- 436.7 W
Get started
Original model resources
Base model resources may not match this quantization. Use the verified deployment weights above when available.
Sources & evidence
1 measured records in total, each traceable to its original source
v0.24 · Ubuntu 22.04 · CUDA 12.8 · 8192 ctx
248.25 tok/s
Decode
794.5 tok/s
Prefill
1.42 s
TTFT
30.4 GB
VRAM
—
MTP
436.7 W
Power draw
L2 batch seed 2026-10-10: vLLM on hw_id=13 running model_id=5 at Q2_K