EXL3 models
Collection
73 items • Updated • 87
EXL3 quants of deepseek-ai/DeepSeek-V4-Flash-0731
⚠️ Needs ExLlamav3 v1.4.0 (or dev branch)
2.04 bits per weight
2.52 bits per weight
2.77 bits per weight
3.04 bits per weight
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support