EXL3 quantization of shisa-v2.1-qwen3-8b, 4 bits per weight.

HumanEval (argmax)

Model Q4 Q6 Q8 FP16
shisa-v2.1-qwen3-8b-exl3-4bpw 82.9 84.8 84.8 84.1
shisa-v2.1-qwen3-8b-exl3-6bpw 86.0 84.8 86.6 86.6
shisa-v2.1-qwen3-8b-exl3-8bpw-h8 84.8 87.8 86.6 86.0
Qwen3-8B-exl3-4bpw 86.0 85.4 86.0 87.2
Qwen3-8B-exl3-6bpw 84.8 86.0 87.2 87.2
Qwen3-8B-exl3-8bpw-h8 86.0 87.2 86.6 86.6
Downloads last month
9
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for isogen/shisa-v2.1-qwen3-8b-exl3-4bpw

Base model

Qwen/Qwen3-8B-Base
Finetuned
Qwen/Qwen3-8B
Quantized
(9)
this model