Back to search results
Qwen2.5-0.5B-Instruct · IQ2_M
Apple M5 Max128GBllama.cpp v0.23L2 Cross-framework verified
This page aggregates 1 real-world runs of Qwen2.5-0.5B-Instruct (IQ2_M) on Apple M5 Max with llama.cpp, contributed by 1 independent source platforms; metrics are averages of published measurements.
298.63 tok/s
Decode
Decode speed
1,141.07 tok/s
Prefill
Prefill speed
2.24 s
TTFT
Time to first token
2.65 GB
VRAM
VRAM usage
L2 Cross-framework verified
1
Measured runs
1
Independent sources
Other
Source platforms
5 days ago
Last verified
Performance
- vLLM · IQ2_XS Decode 291.82 · Prefill 1370.02 ·
- Strata · IQ1_S Decode 343.91 · Prefill 1513.7 ·
- Strata · Q5_K_S Decode 305.21 · Prefill 1469.7 ·
- llama.cpp · IQ2_M (current)Decode 298.63 · Prefill 1141.07 ·
- llama.cpp · Q8_0 Decode 366.43 · Prefill 1208.56 ·
- vLLM · Q2_K Decode 318.68 · Prefill 1287.78 ·
Core figures
- Decode (avg)
- 298.63 tok/s
- Prefill (avg)
- 1,141.07 tok/s
- TTFT (avg)
- 2.24 s
- VRAM (avg)
- 2.65 GB
- MTP acceptance rate
- —
- TTFB
- 2.39 GB
- Power draw
- 73.7 W
Configuration
Member-level fields are taken from the most recent run
- Model
- Qwen2.5-0.5B-Instruct
- Quantization
- IQ2_M
- Framework
- llama.cpp
- Version
- v0.23
- Context length
- 16384 tokens
- Batch size
- 512
- GPU layers
- 20
- 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)
- 2.65 GB
- OS
- macOS 15.0
- Driver
- Apple M5 Max
- CUDA
- N/A
- Power draw
- 73.7 W
Sources & evidence
1 measured records in total, each traceable to its original source
v0.23 · macOS 15.0 · CUDA N/A · 16384 ctx
298.63 tok/s
Decode
1,141.07 tok/s
Prefill
2.24 s
TTFT
2.65 GB
VRAM
—
MTP
73.7 W
Power draw
L2 batch seed 2026-10-13: llama.cpp on hw_id=28 running model_id=59 at IQ2_M