Back to search results
Qwen3-Coder-Next · Q5_K_M
Apple M2 Prollama.cpp v0.23L2 Cross-framework verified
This page aggregates 1 real-world runs of Qwen3-Coder-Next (Q5_K_M) on Apple M2 Pro with llama.cpp, contributed by 1 independent source platforms; metrics are averages of published measurements.
54.45 tok/s
Decode
Decode speed
244.86 tok/s
Prefill
Prefill speed
1.51 s
TTFT
Time to first token
1.05 GB
VRAM
VRAM usage
L2 Cross-framework verified
1
Measured runs
1
Independent sources
Other
Source platforms
2 days ago
Last verified
Performance
- llama.cpp · Q5_K_M (current)Decode 54.45 · Prefill 244.86 ·
- vLLM · Q3_K_S Decode 89.59 · Prefill 433.4 ·
Core figures
- Decode (avg)
- 54.45 tok/s
- Prefill (avg)
- 244.86 tok/s
- TTFT (avg)
- 1.51 s
- VRAM (avg)
- 1.05 GB
- MTP acceptance rate
- —
- TTFB
- 0.44 GB
- Power draw
- 28.9 W
Configuration
Member-level fields are taken from the most recent run
- Model
- Qwen3-Coder-Next
- Quantization
- Q5_K_M
- Framework
- llama.cpp
- Version
- v0.23
- Context length
- 2048 tokens
- Batch size
- 256
- GPU layers
- 20
- 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)
- 1.05 GB
- OS
- macOS 15.0
- Driver
- Apple M2 Pro
- CUDA
- N/A
- Power draw
- 28.9 W
Sources & evidence
1 measured records in total, each traceable to its original source
v0.23 · macOS 15.0 · CUDA N/A · 2048 ctx
54.45 tok/s
Decode
244.86 tok/s
Prefill
1.51 s
TTFT
1.05 GB
VRAM
—
MTP
28.9 W
Power draw
L2 batch seed 2026-10-16: llama.cpp on hw_id=102 running model_id=58 at Q5_K_M