Qwen3-Next-80B-A3B Quantized Models
Collection
FP8-dynamic, FP8-block, NVFP4, INT4, INT8 versions of Qwen3-Next-80B-A3B-Instruct and Qwen3-Next-80B-A3B-Thinking Models
•
8 items
•
Updated
Accuracy
| Category | Metric | Qwen3-Next-80B-A3B-Instruct | Qwen3-Next-80B-A3B-Instruct-quantized.w4a16 | Recovery (%) |
|---|---|---|---|---|
| OpenLLM V1 | ARC-Challenge (Acc-Norm, 25-shot) | 75.85 | 74.83 | 98.65 |
| GSM8K (Strict-Match, 5-shot) | 31.01 | 28.05 | 90.46 | |
| HellaSwag (Acc-Norm, 10-shot) | 83.25 | 83.08 | 99.80 | |
| MMLU (Acc, 5-shot) | 85.56 | 85.03 | 99.38 | |
| TruthfulQA (MC2, 0-shot) | 60.70 | 60.63 | 99.88 | |
| Winogrande (Acc, 5-shot) | 78.30 | 78.37 | 100.10 | |
| Average Score | 69.11 | 68.33 | 98.87 | |
| OpenLLM V2 | IFEval (Inst Level Strict Acc, 0-shot) | 90.41 | 91.37 | 101.06 |
| BBH (Acc-Norm, 3-shot) | 67.78 | 67.33 | 99.33 | |
| Math-Hard (Exact-Match, 4-shot) | 56.04 | 55.36 | 98.79 | |
| GPQA (Acc-Norm, 0-shot) | 28.61 | 28.61 | 100.00 | |
| MUSR (Acc-Norm, 0-shot) | 39.68 | 40.08 | 101.00 | |
| MMLU-Pro (Acc, 5-shot) | 59.73 | 58.37 | 97.72 | |
| Average Score | 57.04 | 56.85 | 99.67 |
- Downloads last month
- 46
Inference Providers
NEW
This model isn't deployed by any Inference Provider.
🙋
Ask for provider support