Back to search results
Qwen2.5-0.5B-Instruct · Q2_K
Apple M1llama.cpp v0.39L2 Cross-framework verified
This page aggregates 1 real-world runs of Qwen2.5-0.5B-Instruct (Q2_K) on Apple M1 with llama.cpp, contributed by 1 independent source platforms; metrics are averages of published measurements.
189.43 tok/s
Decode
Decode speed
860.24 tok/s
Prefill
Prefill speed
1.36 s
TTFT
Time to first token
2.68 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 · IQ2_M Decode 189.83 · Prefill 1001.05 ·
- llama.cpp · Q2_K (current)Decode 189.43 · Prefill 860.24 ·
- vLLM · Q3_K_S Decode 170.33 · Prefill 911.66 ·
- vLLM · Q4_K_M Decode 143.97 · Prefill 610.72 ·
Core figures
- Decode (avg)
- 189.43 tok/s
- Prefill (avg)
- 860.24 tok/s
- TTFT (avg)
- 1.36 s
- VRAM (avg)
- 2.68 GB
- MTP acceptance rate
- —
- TTFB
- 1.52 GB
- Power draw
- 50.9 W
Configuration
Member-level fields are taken from the most recent run
- Model
- Qwen2.5-0.5B-Instruct
- Quantization
- Q2_K
- Framework
- llama.cpp
- Version
- v0.39
- Context length
- 4096 tokens
- Batch size
- 512
- 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 M1
- Measured VRAM (avg)
- 2.68 GB
- OS
- macOS 15.0
- Driver
- Apple M1
- CUDA
- N/A
- Power draw
- 50.9 W
Sources & evidence
1 measured records in total, each traceable to its original source
v0.39 · macOS 15.0 · CUDA N/A · 4096 ctx
189.43 tok/s
Decode
860.24 tok/s
Prefill
1.36 s
TTFT
2.68 GB
VRAM
—
MTP
50.9 W
Power draw
L2 batch seed 2026-10-16: llama.cpp on hw_id=99 running model_id=59 at Q2_K