Back to search results
Qwen3-4B · IQ1_S
Intel Arc A770 16GB16GBllama.cpp v0.53L2 Cross-framework verified
This page aggregates 1 real-world runs of Qwen3-4B (IQ1_S) on Intel Arc A770 16GB with llama.cpp, contributed by 1 independent source platforms; metrics are averages of published measurements.
92.68 tok/s
Decode
Decode speed
463.17 tok/s
Prefill
Prefill speed
2.23 s
TTFT
Time to first token
5.63 GB
VRAM
VRAM usage
L2 Cross-framework verified
1
Measured runs
1
Independent sources
Other
Source platforms
5 days ago
Last verified
Performance
- vLLM · IQ4_XS Decode 72.03 · Prefill 234.09 ·
- Strata · Q6_K Decode 68.88 · Prefill 245.03 ·
- llama.cpp · IQ1_S (current)Decode 92.68 · Prefill 463.17 ·
- vLLM · Q5_K_S Decode 59.1 · Prefill 316.05 ·
Core figures
- Decode (avg)
- 92.68 tok/s
- Prefill (avg)
- 463.17 tok/s
- TTFT (avg)
- 2.23 s
- VRAM (avg)
- 5.63 GB
- MTP acceptance rate
- —
- TTFB
- 0.74 GB
- Power draw
- 240.3 W
Configuration
Member-level fields are taken from the most recent run
- Model
- Qwen3-4B
- Quantization
- IQ1_S
- Framework
- llama.cpp
- Version
- v0.53
- Context length
- 32768 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)
- 5.63 GB
- OS
- Ubuntu 24.04
- Driver
- Intel Arc 31.0.101.5186
- CUDA
- 12.6
- Power draw
- 240.3 W
Sources & evidence
1 measured records in total, each traceable to its original source
v0.53 · Ubuntu 24.04 · CUDA 12.6 · 32768 ctx
92.68 tok/s
Decode
463.17 tok/s
Prefill
2.23 s
TTFT
5.63 GB
VRAM
—
MTP
240.3 W
Power draw
L2 batch seed 2026-10-13: llama.cpp on hw_id=97 running model_id=54 at IQ1_S