EXL3 quants of Qwen3.5-35B-A3B

⚠️ Requires ExLlamaV3 v0.0.23 (or v0.0.22 dev branch)

Base bitrates:

2.00 bits per weight
3.00 bits per weight
4.00 bits per weight
5.00 bits per weight
6.00 bits per weight

Optimized:

2.13 bits per weight
2.54 bits per weight
3.09 bits per weight
3.54 bits per weight
4.09 bits per weight

Benchmarks

kld

SVG Catbench

2.00 bpw
2.00 bpw
2.13 bpw
2.13 bpw
2.54 bpw
2.54 bpw
3.00 bpw
3.00 bpw
3.09 bpw
3.09 bpw
3.54 bpw
3.54 bpw
4.00 bpw
4.00 bpw
4.09 bpw
4.09 bpw
5.00 bpw
5.00 bpw
6.00 bpw
6.00 bpw
FP16
FP16

(more coming soon)

Downloads last month

-

Downloads are not tracked for this model. How to track
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for turboderp/Qwen3.5-35B-A3B-exl3

Quantized
(82)
this model

Collection including turboderp/Qwen3.5-35B-A3B-exl3