MLX 8-bit conversion of Nex-N2.5-mini, a sparse MoE language model for local inference, coding, reasoning, and long-context work. The source checkpoint supports a native context length of 262,144 tokens (256K).
Benchmark results reported by Nex AI for the original Nex-N2.5 checkpoint and its upstream evaluation setup.
| Format | Quantization | Size |
|---|---|---|
| MLX safetensors | Affine 8-bit, group size 64 | 36.85 GB |
This release contains the text-generation weights and tokenizer. It does not include MTP weights or a vision projector.
pip install -U mlx-lm
mlx_lm.generate \
--model . \
--prompt "Explain why reproducible builds matter." \
--max-tokens 512
model-00003-of-00007.safetensors 5.4 GBmodel-00005-of-00007.safetensors 5.4 GBmodel-00004-of-00007.safetensors 5.4 GBmodel-00006-of-00007.safetensors 5.4 GBmodel-00002-of-00007.safetensors 5.4 GBmodel-00001-of-00007.safetensors 5.3 GBmodel-00007-of-00007.safetensors 4.7 GB有文件？提交种子。
提交种子nex-agi/Nex-N2.5-mini