Ornith-1.5-9B-oQ4e-fp16

Updated 2026-11-19: Re-quantized with oMLX v0.8.0. Please re-download if you have an earlier copy.

This model was quantized using oQ (oMLX v0.8.0) mixed-precision quantization.

The effective average is 6.143 bits/weight for the complete checkpoint.

Quantization and Bit Distribution

Storage format Logical weights Share of total Tensor storage Effective bits/weight
Affine 4-bit, group 64 6.201B 65.902% 3.249 GiB 4.50
Affine 5-bit, group 64 2.592B 27.541% 1.659 GiB 5.50
Affine 6-bit, group 64 0.160B 1.697% 0.121 GiB 6.50
Affine 8-bit, group 64 0.000B 0.003% 0.000 GiB 8.50
FP16 0.001B 0.011% 0.002 GiB 16.00
FP32 0.456B 4.846% 1.699 GiB 32.00
Total 9.410B 100% 6.730 GiB 6.143

Storage Breakdown

Component Logical weights Storage
Language model 8.954B 5.402 GB / 5.031 GiB
Vision 0.456B 1.824 GB / 1.699 GiB
Total 9.410B 7.226 GB / 6.730 GiB
Downloads last month
302
Safetensors
Model size
9B params
Tensor type
U32
·
F16
·
F32
·
MLX
Hardware compatibility
Log In to add your hardware

4-bit

Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support