Qwen3.8-27B · Q3_K_L
This page aggregates 1 real-world runs of Qwen3.8-27B (Q3_K_L) on Apple M5 Max with llama.cpp, contributed by 1 independent source platforms; metrics are averages of published measurements.
42.74 tok/s
Decode
Decode speed
158.84 tok/s
Prefill
Prefill speed
1.45 s
TTFT
Time to first token
42.92 GB
VRAM
VRAM usage
1
Measured runs
1
Independent sources
Other
Source platforms
16 days ago
Last verified
Performance
- llama.cpp · Q3_K_L (current)Decode 42.74 · Prefill 158.84 ·
- llama.cpp · Q6_K Decode 30.73 · Prefill 132.88 ·
- Ollama · IQ3_XS Decode 46.88 · Prefill 178.13 ·
- llama.cpp · Q5_K_M Decode 23.9 · Prefill 179.3 ·
Core figures
- Decode (avg)
- 42.74 tok/s
- Prefill (avg)
- 158.84 tok/s
- TTFT (avg)
- 1.45 s
- VRAM (avg)
- 42.92 GB
- MTP acceptance rate
- —
- TTFB
- 0.34 GB
- Power draw
- 54.1 W
Configuration
Member-level fields are taken from the most recent run
- Model
- Qwen3.8-27B
- Quantization
- Q3_K_L
- Framework
- llama.cpp
- Version
- v0.46
- Context length
- 4096 tokens
- Batch size
- 512
- GPU layers
- 28
- Flash Attention
- On
Hardware
Nominal and measured figures are shown side by side; whether it runs is the reader's call
- GPU
- Apple M5 Max
- Nominal VRAM
- 128 GB
- Measured VRAM (avg)
- 42.92 GB
- OS
- macOS 26.1
- Power draw
- 54.1 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
v0.46 · macOS 26.1 · 4096 ctx
42.74 tok/s
Decode
158.84 tok/s
Prefill
1.45 s
TTFT
42.92 GB
VRAM
—
MTP
54.1 W
Power draw
L2 batch seed 2026-10-19: llama.cpp on hw_id=28 running model_id=5 at Q3_K_L