LLM EXPLORER 56,604 MODELS INDEXED

Inkling Small MLX REAP25 4bit by pipenetwork

By pipenetwork · 312 downloads
HF Repo ↗ Copy URL

Inkling Small MLX REAP25 4bit is an open-source language model by pipenetwork. Features: 31.2b LLM, VRAM: 112.3GB, License: apache-2.0, Quantized.

  Arxiv:2510.13999   4bit   Audio-text-to-text Base model:quantized:thinkingm... Base model:thinkingmachines/in...   Conversational   Image-text-to-text   Inkling   Inkling mm model   Mlx   Moe   Multimodal   Pruned   Quantized   Reap   Region:us   Safetensors   Sharded   Tensorflow   Thinking-machines
Model Card on HF 🤗: https://huggingface.co/pipenetwork/Inkling-Small-MLX-REAP25-4bit 

Inkling Small MLX REAP25 4bit Parameters and Internals

LLM NameInkling Small MLX REAP25 4bit
Repository 🤗https://huggingface.co/pipenetwork/Inkling-Small-MLX-REAP25-4bit 
Base Model(s)  Inkling Small   thinkingmachines/Inkling-Small
Model Size31.2b
Required VRAM112.3 GB
Updated2026-08-09
Maintainerpipenetwork
Model Typeinkling_mm_model
Model Files  5.4 GB: 1-of-21   5.5 GB: 2-of-21   5.5 GB: 3-of-21   5.7 GB: 4-of-21   5.7 GB: 5-of-21   5.6 GB: 6-of-21   5.6 GB: 7-of-21   5.7 GB: 8-of-21   5.5 GB: 9-of-21   5.5 GB: 10-of-21   5.7 GB: 11-of-21   5.0 GB: 12-of-21   5.7 GB: 13-of-21   5.6 GB: 14-of-21   5.5 GB: 15-of-21   5.6 GB: 16-of-21   5.5 GB: 17-of-21   5.6 GB: 18-of-21   5.0 GB: 19-of-21   5.4 GB: 20-of-21   2.0 GB: 21-of-21
Quantization Type4bit
Model ArchitectureInklingForConditionalGeneration
Licenseapache-2.0
Tokenizer ClassPreTrainedTokenizerFast