---
license: apache-2.0
base_model: Nanbeige/Nanbeige4.2-3B
language: [en, zh]
library_name: mlx
pipeline_tag: text-generation
tags: [mlx, nanbeige, looped-transformer, apple-silicon, 6-bit]
quantized_from: Nanbeige/Nanbeige4.2-3B
quantization:
  bits: 6
  group_size: 64
---

# Nanbeige4.2-3B MLX (6-bit)

An MLX conversion and 6-bit quantization (group size 64) of
[`Nanbeige/Nanbeige4.2-3B`](https://huggingface.co/Nanbeige/Nanbeige4.2-3B) — a
3-billion-parameter **Looped Transformer** (`num_loops=2`, weight-shared over 22
layers for an effective depth of 44). Produced by the independent
[`nanbeige-mlx`](https://github.com/jishnuvenugopal/nanbeige-mlx) project;
not affiliated with the Nanbeige team.

## Load (no conversion step)

```python
from nanbeige_mlx import pull
import mlx_lm
model, tok = mlx_lm.load(pull("6bit"))
```

Weights arrive on first use and are cached in `~/.cache/huggingface`.

## KV-cache ceiling (a real cost of the looped design)

The looped architecture needs `num_loops * num_hidden_layers = 44` KV slots.
At full context that is 44 x 8 KV-heads x 128 dim x 2 (K+V) x 262144 positions
x 2 bytes ~= **47 GB** — unreachable on a 16 GB machine. Because this model
supplies `make_cache`, mlx-lm's `--max-kv-size` knob is inert; use `--kv-bits`
to reduce KV precision instead.

## Fidelity of the port

This quant is produced by a from-scratch MLX port of the looped architecture,
checked against the HuggingFace reference. The headline finding carries over
verbatim from the port's documentation, and must read identically wherever it
appears (repo README, model card, investigation log):

> Bottom line: **the source of the logit gap is documented-open, not blocking.**
> Six candidate causes are ruled out by measurement (the seventh, "RoPE runs in
> bf16," was killed by a bit-identical upcast experiment — `mx.fast.rope` is
> fp32-internal regardless of input dtype). The `[1,1,5,6]` mask shape is a red
> herring: the 6th column is HF's `past_seen_tokens + sequence_length + 1`
> boilerplate for StaticCache sizing, present in every Llama-family model, and
> eager attention slices it to `[:,:,:,:L]` before use — nothing to do with
> `num_loops`. The gap does not affect behavior on the agentic suite. Bit-exact
> parity is not claimed and not expected in bf16. Full record:
> [`docs/investigation-log.md`](https://github.com/jishnuvenugopal/nanbeige-mlx-eval/blob/main/docs/investigation-log.md).

## License

Apache-2.0 (the upstream license; quantization is a modification under §4(d)).
The `nanbeige.py` model definition is MIT-licensed source from `nanbeige-mlx`.
See `LICENSE` and `NOTICE`.