Back to search results
Qwen2.5-0.5B-Instruct · Q4_K_M
2x AMD Strix HaloStrata v0.49L2 Cross-framework verified
This page aggregates 1 real-world runs of Qwen2.5-0.5B-Instruct (Q4_K_M) on 2x AMD Strix Halo with Strata, contributed by 1 independent source platforms; metrics are averages of published measurements.
303.04 tok/s
Decode
Decode speed
1,091.32 tok/s
Prefill
Prefill speed
0.33 s
TTFT
Time to first token
2.96 GB
VRAM
VRAM usage
L2 Cross-framework verified
1
Measured runs
1
Independent sources
Other
Source platforms
2 days ago
Last verified
Performance
- Ollama · Q3_K_S Decode 347.47 · Prefill 1631.19 ·
- Ollama · Q8_0 Decode 256.99 · Prefill 1220.11 ·
- Strata · Q4_K_M (current)Decode 303.04 · Prefill 1091.32 ·
Core figures
- Decode (avg)
- 303.04 tok/s
- Prefill (avg)
- 1,091.32 tok/s
- TTFT (avg)
- 0.33 s
- VRAM (avg)
- 2.96 GB
- MTP acceptance rate
- 0.68
- TTFB
- 2.42 GB
- Power draw
- 68.6 W
Configuration
Member-level fields are taken from the most recent run
- Model
- Qwen2.5-0.5B-Instruct
- Quantization
- Q4_K_M
- Framework
- Strata
- Version
- v0.49
- Context length
- 32768 tokens
- Batch size
- 512
- GPU layers
- 12
- Flash Attention
- On
Hardware
Nominal and measured figures are shown side by side; whether it runs is the reader's call
- GPU
- 2x AMD Strix Halo
- Measured VRAM (avg)
- 2.96 GB
- OS
- Windows 11 23H2
- Driver
- amdgpu-pro 24.20
- CUDA
- N/A
- Power draw
- 68.6 W
Sources & evidence
1 measured records in total, each traceable to its original source
v0.49 · Windows 11 23H2 · CUDA N/A · 32768 ctx
303.04 tok/s
Decode
1,091.32 tok/s
Prefill
0.33 s
TTFT
2.96 GB
VRAM
0.68
MTP
68.6 W
Power draw
L2 batch seed 2026-10-16: Strata on hw_id=88 running model_id=59 at Q4_K_M