Back to search results
Qwen3-4B-Thinking-2507 · IQ2_M
2x AMD Strix HaloStrata v0.54L2 Cross-framework verified
This page aggregates 1 real-world runs of Qwen3-4B-Thinking-2507 (IQ2_M) on 2x AMD Strix Halo with Strata, contributed by 1 independent source platforms; metrics are averages of published measurements.
312.05 tok/s
Decode
Decode speed
1,568.66 tok/s
Prefill
Prefill speed
0.71 s
TTFT
Time to first token
6.47 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 · Q3_K_M Decode 181.31 · Prefill 944.53 ·
- llama.cpp · Q4_K_M Decode 153.87 · Prefill 676.15 ·
- Strata · IQ2_M (current)Decode 312.05 · Prefill 1568.66 ·
Core figures
- Decode (avg)
- 312.05 tok/s
- Prefill (avg)
- 1,568.66 tok/s
- TTFT (avg)
- 0.71 s
- VRAM (avg)
- 6.47 GB
- MTP acceptance rate
- 0.65
- TTFB
- 2.5 GB
- Power draw
- 46 W
Configuration
Member-level fields are taken from the most recent run
- Model
- Qwen3-4B-Thinking-2507
- Quantization
- IQ2_M
- Framework
- Strata
- Version
- v0.54
- Context length
- 32768 tokens
- Batch size
- 256
- GPU layers
- 5
- 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)
- 6.47 GB
- OS
- Ubuntu 24.04
- Driver
- amdgpu-pro 24.20
- CUDA
- N/A
- Power draw
- 46 W
Sources & evidence
1 measured records in total, each traceable to its original source
v0.54 · Ubuntu 24.04 · CUDA N/A · 32768 ctx
312.05 tok/s
Decode
1,568.66 tok/s
Prefill
0.71 s
TTFT
6.47 GB
VRAM
0.65
MTP
46 W
Power draw
L2 batch seed 2026-10-16: Strata on hw_id=88 running model_id=57 at IQ2_M