Qwen3.6-35B-A3B · Q8_0
This page aggregates 1 real-world runs of Qwen3.6-35B-A3B (Q8_0) on NVIDIA RTX Spark with llama.cpp, contributed by 1 independent source platforms; metrics are averages of published measurements.
57.81 tok/s
Decode
Decode speed
190.18 tok/s
Prefill
Prefill speed
3.49 s
TTFT
Time to first token
86.62 GB
VRAM
VRAM usage
1
Measured runs
1
Independent sources
Other
Source platforms
16 days ago
Last verified
Performance
- llama.cpp · Q5_K_M Decode 78.77 · Prefill 418.38 ·
- llama.cpp · Q8_0 (current)Decode 57.81 · Prefill 190.18 ·
- Ollama · Q8_0 Decode 52.36 · Prefill 271.04 ·
- llama.cpp · IQ3_XS Decode 117.69 · Prefill 448.38 ·
- llama.cpp · Q4_K_M Decode 88.7 · Prefill 354.07 ·
- Ollama · IQ2_XS Decode 128.3 · Prefill 530.41 ·
- Ollama · Q4_K_M Decode 85.36 · Prefill 466.28 ·
Core figures
- Decode (avg)
- 57.81 tok/s
- Prefill (avg)
- 190.18 tok/s
- TTFT (avg)
- 3.49 s
- VRAM (avg)
- 86.62 GB
- MTP acceptance rate
- —
- TTFB
- 0.46 GB
- Power draw
- 57.2 W
Configuration
Member-level fields are taken from the most recent run
- Model
- Qwen3.6-35B-A3B
- Quantization
- Q8_0
- Framework
- llama.cpp
- Version
- v0.54
- 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)
- 86.62 GB
- OS
- Ubuntu 24.04
- Driver
- 615.78
- CUDA
- 13.0
- Power draw
- 57.2 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
v0.54 · Ubuntu 24.04 · CUDA 13.0 · 8192 ctx
57.81 tok/s
Decode
190.18 tok/s
Prefill
3.49 s
TTFT
86.62 GB
VRAM
—
MTP
57.2 W
Power draw
L2 batch seed 2026-10-20 (Linux 116GiB unlock): llama.cpp on Spark running model_id=47 at Q8_0