Arao AI
Back to search results

Qwen3.5-27B · Q2_K

Start a discussion
NVIDIA RTX 509032GBStrata v0.31L2 Cross-framework verified

This page aggregates 1 real-world runs of Qwen3.5-27B (Q2_K) on NVIDIA RTX 5090 with Strata, contributed by 1 independent source platforms; metrics are averages of published measurements.

366.75 tok/s

Decode

Decode speed

1,325.45 tok/s

Prefill

Prefill speed

2.28 s

TTFT

Time to first token

30.4 GB

VRAM

VRAM usage

L2 Cross-framework verified

1

Measured runs

1

Independent sources

Other

Source platforms

Today

Last verified

Performance

  1. Strata · Q2_K (current)Decode 366.75 · Prefill 1325.45 ·
  2. Ollama · IQ1_S Decode 260.56 · Prefill 868.47 ·
  3. Ollama · IQ2_M Decode 235.26 · Prefill 1229.12 ·
  4. llama.cpp · IQ4_XS Decode 134 · Prefill 1478 ·
  5. llama.cpp · Q5_K_M Decode 114 · Prefill 1331 ·
  6. llama.cpp · Q8_0 Decode 92 · Prefill 792 ·
  7. vLLM · IQ4_XS Decode 108 · Prefill 1288 ·
  8. vLLM · Q4_K_M Decode 100 · Prefill 1182 ·
  9. vLLM · Q5_K_M Decode 106 · Prefill 956 ·
  10. vLLM · Q8_0 Decode 74 · Prefill 795 ·
  11. Ollama · IQ4_XS Decode 136 · Prefill 1332 ·
  12. Ollama · Q4_K_M Decode 125 · Prefill 1177 ·
  13. Ollama · Q5_K_M Decode 113 · Prefill 1178 ·
  14. Ollama · Q8_0 Decode 75 · Prefill 899 ·
  15. Strata · IQ4_XS Decode 198 · Prefill 1923 ·
  16. Strata · Q4_K_M Decode 178 · Prefill 1794 ·
  17. Strata · Q5_K_M Decode 147 · Prefill 1764 ·
  18. Strata · Q8_0 Decode 127 · Prefill 1290 ·
  19. llama.cpp · Q4_K_M Decode 30.09 · Prefill 1515.83 ·

Core figures

Decode (avg)
366.75 tok/s
Prefill (avg)
1,325.45 tok/s
TTFT (avg)
2.28 s
VRAM (avg)
30.4 GB
MTP acceptance rate
0.63
TTFB
2.93 GB
Power draw
420 W

Configuration

Member-level fields are taken from the most recent run

Model
Qwen3.5-27B
Quantization
Q2_K
Framework
Strata
Version
v0.31
Context length
32768 tokens
Batch size
512
GPU layers
15
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
560.94
CUDA
12.4
Power draw
420 W

Sources & evidence

1 measured records in total, each traceable to its original source

  1. L2 Cross-framework verifiedOtherOriginal link Verified on 2026-10-18

    v0.31 · Ubuntu 22.04 · CUDA 12.4 · 32768 ctx

    366.75 tok/s

    Decode

    1,325.45 tok/s

    Prefill

    2.28 s

    TTFT

    30.4 GB

    VRAM

    0.63

    MTP

    420 W

    Power draw

    L2 batch seed 2026-10-18: Strata on hw_id=13 running model_id=50 at Q2_K