gemma-4-31B-it-oQ4e-mtp
[!NOTE] Updated 2026-10-29: Re-quantized with the improved oQe quantization in oMLX 0.7.3. Please re-download if you have an earlier copy.
This model was quantized using oQ (oMLX v0.7.3) mixed-precision quantization with improved oQe (jundot/omlx#4930).
A mixed-precision MLX checkpoint of google/gemma-4-31B-it, the 30.7B dense model of the Gemma 4 family from Google DeepMind, with text and image input, configurable thinking modes and a 256K-token context. It includes the language model, the vision encoder and a multi-token prediction (MTP) head merged from google/gemma-4-31B-it-assistant.
The effective average is 4.801 bits/weight for the language backbone and 5.167 bits/weight for the complete checkpoint.
Quantization and Bit Distribution
oQ4e allocates bits by measured layer sensitivity, then rounds each group with an importance matrix (128 samples x 512 tokens) and a weighted least-squares refit of its scale and bias.
| Storage format | Logical weights | Share of total | Tensor storage | Effective bits/weight |
|---|---|---|---|---|
| Affine 4-bit, group 64 | 21.790B | 68.645% | 11.415 GiB | 4.50 |
| Affine 5-bit, group 64 | 8.615B | 27.141% | 5.516 GiB | 5.50 |
| Affine 6-bit, group 64 | 0.297B | 0.937% | 0.225 GiB | 6.50 |
| BF16 | 1.040B | 3.278% | 1.938 GiB | 16.00 |
| Total | 31.743B | 100% | 19.094 GiB | 5.167 |
Storage Breakdown
| Component | Logical weights | Storage |
|---|---|---|
| Language backbone | 30.697B | 18.420 GB / 17.155 GiB |
| Vision encoder | 0.576B | 1.143 GB / 1.064 GiB |
| MTP head | 0.470B | 0.939 GB / 0.875 GiB |
| Total | 31.743B | 20.502 GB / 19.094 GiB |
License
Apache-2.0, following the source model. See the Gemma 4 license.