Language Model Evaluation on Qwen3-0.6B Evaluation Suite (average)
47.8Average PerformanceEDGERAZOR
Evaluation Results
| Method | Links | |
|---|---|---|
| EDGERAZORQuantization Configuration (W-A-KV)=W4-A8-KV82026.04 | 47.8 | |
| BF16Quantization Configuration (W-A-KV)=W16-A16-KV162026.04 | 47.35 | |
| FlatQuantQuantization Configuration (W-A-KV)=W4-A8-KV82026.04 | 45.74 | |
| LQERQuantization Configuration (W-A-KV)=W4-A8-KV82026.04 | 45.31 | |
| ABQ-LLMQuantization Configuration (W-A-KV)=W4-A8-KV82026.04 | 44.52 | |
| EDGERAZORQuantization Configuration (W-A-KV)=W3-A8-KV82026.04 | 44.1 | |
| EDGERAZORQuantization Configuration (W-A-KV)=W2-A8-KV82026.04 | 41.76 | |
| SpinQuantQuantization Configuration (W-A-KV)=W4-A8-KV82026.04 | 41.27 | |
| EDGERAZORQuantization Configuration (W-A-KV)=W1.58-A8-KV82026.04 | 39.81 | |
| FlatQuantQuantization Configuration (W-A-KV)=W3-A8-KV82026.04 | 37.38 | |
| OmniQuantQuantization Configuration (W-A-KV)=W4-A8-KV82026.04 | 37.27 | |
| LQERQuantization Configuration (W-A-KV)=W3-A8-KV82026.04 | 36.46 | |
| SpinQuantQuantization Configuration (W-A-KV)=W3-A8-KV82026.04 | 34.93 | |
| OmniQuantQuantization Configuration (W-A-KV)=W3-A8-KV82026.04 | 34.58 | |
| ABQ-LLMQuantization Configuration (W-A-KV)=W3-A8-KV82026.04 | 31.72 | |
| OmniQuantQuantization Configuration (W-A-KV)=W2-A8-KV82026.04 | 30.49 | |
| LQERQuantization Configuration (W-A-KV)=W2-A8-KV82026.04 | 30.46 | |
| ABQ-LLMQuantization Configuration (W-A-KV)=W2-A8-KV82026.04 | 30.4 | |
| FlatQuantQuantization Configuration (W-A-KV)=W2-A8-KV82026.04 | 30.23 | |
| QuaRotQuantization Configuration (W-A-KV)=W4-A8-KV82026.04 | 30.12 | |
| QuaRotQuantization Configuration (W-A-KV)=W2-A8-KV82026.04 | 30.12 | |
| SpinQuantQuantization Configuration (W-A-KV)=W2-A8-KV82026.04 | 30.04 | |
| QuaRotQuantization Configuration (W-A-KV)=W3-A8-KV82026.04 | 29.81 | |
| QoQQuantization Configuration (W-A-KV)=W4-A8-KV82026.04 | 29.8 |