MXFP8 is a block floating-point format supported by MLX on Apple Silicon. All eligible linear weights use MXFP8; the config records the quantization parameters.

About the base model

Zeta 2.1 is a code edit-prediction model (next-edit suggestion) finetuned from ByteDance-Seed/Seed-Coder-8B-Base. It uses an SPM-style, multi-marker FIM format rather than a chat template.

Prompt format

text

<[fim-suffix]>{code after cursor}
<[fim-prefix]><filename>{file_path}
{code before cursor}<|marker_1|><|marker_2|>
<[fim-middle]>

See examples/zeta_fim.py for a standard-library client targeting LM Studio.

Usage

python

from mlx_lm import load, generate
from mlx_lm.sample_utils import make_sampler

model, tokenizer = load("bouroo/zeta-2.1-mlx-mxfp8")
prompt = "<[fim-suffix]>\n    return total\n<[fim-prefix]><filename>app.py\ndef sum_list(items):\n<|marker_1|><|marker_2|>\n<[fim-middle]>"
out = generate(model, tokenizer, prompt=prompt, max_tokens=128,
               sampler=make_sampler(temp=0.2, top_p=0.95))

LM Studio

bash

lms get https://huggingface.co/bouroo/zeta-2.1-mlx-mxfp8
lms load zeta-2.1-mlx-mxfp8

LM Studio's built-in autocomplete uses a simple FIM template and cannot format Zeta's multi-marker edit-prediction prompt. Use the client in examples/ or a compatible editor integration.

Verification

Confirmed with local mlx_lm load + generation using the canonical FIM prompt. The output produced a complete sum_list implementation enclosed by <|marker_1|> and <|marker_2|>.

Quantized and published under Apache-2.0, inherited from the base model.