Back to search results
Qwen2.5-0.5B-Instruct · IQ4_XS
Intel Arc A770 16GB16GBllama.cpp v0.24L2 Cross-framework verified
This page aggregates 1 real-world runs of Qwen2.5-0.5B-Instruct (IQ4_XS) on Intel Arc A770 16GB with llama.cpp, contributed by 1 independent source platforms; metrics are averages of published measurements.
170.02 tok/s
Decode
Decode speed
839.84 tok/s
Prefill
Prefill speed
0.92 s
TTFT
Time to first token
2.88 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 · IQ3_XS Decode 195.32 · Prefill 695.93 ·
- llama.cpp · IQ4_XS (current)Decode 170.02 · Prefill 839.84 ·
- vLLM · IQ1_S Decode 232.59 · Prefill 880.44 ·
- vLLM · IQ2_M Decode 228.63 · Prefill 1047.8 ·
- Strata · Q3_K_M Decode 255.52 · Prefill 902.5 ·
- Strata · Q6_K Decode 176.77 · Prefill 757.14 ·
Core figures
- Decode (avg)
- 170.02 tok/s
- Prefill (avg)
- 839.84 tok/s
- TTFT (avg)
- 0.92 s
- VRAM (avg)
- 2.88 GB
- MTP acceptance rate
- —
- TTFB
- 1.36 GB
- Power draw
- 171.8 W
Configuration
Member-level fields are taken from the most recent run
- Model
- Qwen2.5-0.5B-Instruct
- Quantization
- IQ4_XS
- Framework
- llama.cpp
- Version
- v0.24
- Context length
- 8192 tokens
- Batch size
- 512
- GPU layers
- 33
- Flash Attention
- On
Hardware
Nominal and measured figures are shown side by side; whether it runs is the reader's call
- GPU
- Intel Arc A770 16GB
- Nominal VRAM
- 16 GB
- Measured VRAM (avg)
- 2.88 GB
- OS
- Ubuntu 24.04
- Driver
- Intel Arc 31.0.101.5186
- CUDA
- 12.6
- Power draw
- 171.8 W
Sources & evidence
1 measured records in total, each traceable to its original source
v0.24 · Ubuntu 24.04 · CUDA 12.6 · 8192 ctx
170.02 tok/s
Decode
839.84 tok/s
Prefill
0.92 s
TTFT
2.88 GB
VRAM
—
MTP
171.8 W
Power draw
L2 batch seed 2026-10-16: llama.cpp on hw_id=97 running model_id=59 at IQ4_XS