Qwen3.5-27B · Q8_0
This page aggregates 1 real-world runs of Qwen3.5-27B (Q8_0) on NVIDIA RTX Spark with llama.cpp, contributed by 1 independent source platforms; metrics are averages of published measurements.
6.76 tok/s
Decode
Decode speed
25.64 tok/s
Prefill
Prefill speed
2.76 s
TTFT
Time to first token
67.3 GB
VRAM
VRAM usage
1
Measured runs
1
Independent sources
Other
Source platforms
15 days ago
Last verified
Performance
- llama.cpp · Q6_K Decode 8.03 · Prefill 30.64 ·
- llama.cpp · Q8_0 (current)Decode 6.76 · Prefill 25.64 ·
- Ollama · FP16 Decode 4.9 · Prefill 24.42 ·
- Ollama · Q6_K Decode 7.03 · Prefill 31.25 ·
- Ollama · Q8_0 Decode 6.87 · Prefill 25.86 ·
- llama.cpp · Q4_K_M Decode 10.21 · Prefill 49.89 ·
- Ollama · Q4_K_M Decode 9.77 · Prefill 41.43 ·
Core figures
- Decode (avg)
- 6.76 tok/s
- Prefill (avg)
- 25.64 tok/s
- TTFT (avg)
- 2.76 s
- VRAM (avg)
- 67.3 GB
- MTP acceptance rate
- —
- TTFB
- 0.05 GB
- Power draw
- 50.7 W
Configuration
Member-level fields are taken from the most recent run
- Model
- Qwen3.5-27B
- Quantization
- Q8_0
- Framework
- llama.cpp
- Version
- v0.41
- Context length
- 8192 tokens
- Batch size
- 512
- GPU layers
- 99
- Flash Attention
- On
Hardware
Nominal and measured figures are shown side by side; whether it runs is the reader's call
- GPU
- NVIDIA RTX Spark
- Nominal VRAM
- 128 GB
- Measured VRAM (avg)
- 67.3 GB
- OS
- Ubuntu 24.04
- Driver
- 615.78
- CUDA
- 13.0
- Power draw
- 50.7 W
- Reference price
- $2,399 · updated 2026-10-19
Prices are not live; check the store page
Sources & evidence
1 measured records in total, each traceable to its original source
v0.41 · Ubuntu 24.04 · CUDA 13.0 · 8192 ctx
6.76 tok/s
Decode
25.64 tok/s
Prefill
2.76 s
TTFT
67.3 GB
VRAM
—
MTP
50.7 W
Power draw
L2 batch seed 2026-10-20 (Linux 116GiB unlock): llama.cpp on Spark running model_id=50 at Q8_0