Back to search results
Qwen3.8-9B · Q6_K
Apple M2 ProvLLM v0.31L2 Cross-framework verified
This page aggregates 1 real-world runs of Qwen3.8-9B (Q6_K) on Apple M2 Pro with vLLM, contributed by 1 independent source platforms; metrics are averages of published measurements.
56.64 tok/s
Decode
Decode speed
295.33 tok/s
Prefill
Prefill speed
1.42 s
TTFT
Time to first token
21 GB
VRAM
VRAM usage
L2 Cross-framework verified
1
Measured runs
1
Independent sources
Other
Source platforms
2 days ago
Last verified
Performance
- vLLM · Q5_K_M Decode 58.73 · Prefill 314.79 ·
- vLLM · Q6_K (current)Decode 56.64 · Prefill 295.33 ·
- Ollama · Q8_0 Decode 42.75 · Prefill 231.64 ·
Core figures
- Decode (avg)
- 56.64 tok/s
- Prefill (avg)
- 295.33 tok/s
- TTFT (avg)
- 1.42 s
- VRAM (avg)
- 21 GB
- MTP acceptance rate
- —
- TTFB
- 0.45 GB
- Power draw
- 57.3 W
Configuration
Member-level fields are taken from the most recent run
- Model
- Qwen3.8-9B
- Quantization
- Q6_K
- Framework
- vLLM
- Version
- v0.31
- Context length
- 8192 tokens
- Batch size
- 8
- GPU layers
- 999
- Flash Attention
- Off
Hardware
Nominal and measured figures are shown side by side; whether it runs is the reader's call
- GPU
- Apple M2 Pro
- Measured VRAM (avg)
- 21 GB
- OS
- macOS 15.0
- Driver
- Apple M2 Pro
- CUDA
- N/A
- Power draw
- 57.3 W
Sources & evidence
1 measured records in total, each traceable to its original source
v0.31 · macOS 15.0 · CUDA N/A · 8192 ctx
56.64 tok/s
Decode
295.33 tok/s
Prefill
1.42 s
TTFT
21 GB
VRAM
—
MTP
57.3 W
Power draw
L2 batch seed 2026-10-16: vLLM on hw_id=102 running model_id=52 at Q6_K