Back to search results
Qwen3.8-9B · IQ3_S
AMD Ryzen AI Max+ 395128GBllama.cpp v0.31L2 Cross-framework verified
This page aggregates 1 real-world runs of Qwen3.8-9B (IQ3_S) on AMD Ryzen AI Max+ 395 with llama.cpp, contributed by 1 independent source platforms; metrics are averages of published measurements.
115.85 tok/s
Decode
Decode speed
489.76 tok/s
Prefill
Prefill speed
1.85 s
TTFT
Time to first token
14.57 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 125.51 · Prefill 634.19 ·
- llama.cpp · IQ3_S (current)Decode 115.85 · Prefill 489.76 ·
- vLLM · Q5_K_M Decode 82.89 · Prefill 274.27 ·
- vLLM · IQ2_XS Decode 145.04 · Prefill 523.91 ·
- vLLM · Q3_K_L Decode 107.95 · Prefill 386.71 ·
- Strata · IQ3_XS Decode 164.91 · Prefill 583.87 ·
- Strata · Q5_K_S Decode 114.19 · Prefill 509.48 ·
Core figures
- Decode (avg)
- 115.85 tok/s
- Prefill (avg)
- 489.76 tok/s
- TTFT (avg)
- 1.85 s
- VRAM (avg)
- 14.57 GB
- MTP acceptance rate
- —
- TTFB
- 0.93 GB
- Power draw
- 113 W
Configuration
Member-level fields are taken from the most recent run
- Model
- Qwen3.8-9B
- Quantization
- IQ3_S
- Framework
- llama.cpp
- Version
- v0.31
- 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
- AMD Ryzen AI Max+ 395
- Nominal VRAM
- 128 GB
- Measured VRAM (avg)
- 14.57 GB
- OS
- Ubuntu 24.04
- Driver
- amdgpu 6.10
- CUDA
- N/A
- Power draw
- 113 W
Sources & evidence
1 measured records in total, each traceable to its original source
v0.31 · Ubuntu 24.04 · CUDA N/A · 8192 ctx
115.85 tok/s
Decode
489.76 tok/s
Prefill
1.85 s
TTFT
14.57 GB
VRAM
—
MTP
113 W
Power draw
L2 batch seed 2026-10-16: llama.cpp on hw_id=24 running model_id=52 at IQ3_S