Back to search results
Qwen2.5-Omni-3B · IQ4_XS
AMD Strix Halo (Radeon 8060S)llama.cpp v0.34L2 Cross-framework verified
This page aggregates 1 real-world runs of Qwen2.5-Omni-3B (IQ4_XS) on AMD Strix Halo (Radeon 8060S) with llama.cpp, contributed by 1 independent source platforms; metrics are averages of published measurements.
127.7 tok/s
Decode
Decode speed
503.71 tok/s
Prefill
Prefill speed
0.82 s
TTFT
Time to first token
6.76 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 · IQ4_XS (current)Decode 127.7 · Prefill 503.71 ·
- Ollama · Q4_K_S Decode 115.67 · Prefill 506.03 ·
- Ollama · Q5_K_M Decode 87.89 · Prefill 401.06 ·
Core figures
- Decode (avg)
- 127.7 tok/s
- Prefill (avg)
- 503.71 tok/s
- TTFT (avg)
- 0.82 s
- VRAM (avg)
- 6.76 GB
- MTP acceptance rate
- —
- TTFB
- 1.02 GB
- Power draw
- 58.6 W
Configuration
Member-level fields are taken from the most recent run
- Model
- Qwen2.5-Omni-3B
- Quantization
- IQ4_XS
- Framework
- llama.cpp
- Version
- v0.34
- 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 Strix Halo (Radeon 8060S)
- Measured VRAM (avg)
- 6.76 GB
- OS
- Ubuntu 22.04
- Driver
- amdgpu 6.10
- CUDA
- N/A
- Power draw
- 58.6 W
Sources & evidence
1 measured records in total, each traceable to its original source
v0.34 · Ubuntu 22.04 · CUDA N/A · 8192 ctx
127.7 tok/s
Decode
503.71 tok/s
Prefill
0.82 s
TTFT
6.76 GB
VRAM
—
MTP
58.6 W
Power draw
L2 batch seed 2026-10-16: llama.cpp on hw_id=87 running model_id=60 at IQ4_XS