LLM EXPLORER 59,358 MODELS INDEXED

Inkling NVFP4 Mlx 4bit by mlx-community

By mlx-community · 1119 downloads
HF Repo ↗ Copy URL

Inkling NVFP4 Mlx 4bit is an open-source language model by mlx-community. Features: 176.5b LLM, VRAM: 373.7GB, License: apache-2.0, Quantized, LLM Explorer Score: 0.43, ELO: 1442.

  4bit   Audio Base model:finetune:thinkingma... Base model:thinkingmachines/in...   Conversational   En   Inkling   Inkling mm model   Mlx   Moe   Quantized   Region:us   Safetensors   Sharded   Speech   Tensorflow   Text   Vision
Model Card on HF 🤗: https://huggingface.co/mlx-community/Inkling-NVFP4-mlx-4bit 

Inkling NVFP4 Mlx 4bit Benchmarks

LMSys ELO 1442 vs 1272 (so35) 13.4%
nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").

Inkling NVFP4 Mlx 4bit Parameters and Internals

LLM NameInkling NVFP4 Mlx 4bit
Repository 🤗https://huggingface.co/mlx-community/Inkling-NVFP4-mlx-4bit 
Base Model(s)  Inkling NVFP4   thinkingmachines/Inkling-NVFP4
Model Size176.5b
Required VRAM373.7 GB
Updated2026-08-10
Maintainermlx-community
Model Typeinkling_mm_model
Model Files  5.1 GB: 1-of-73   7.8 GB: 2-of-73   8.2 GB: 3-of-73   8.2 GB: 4-of-73   8.2 GB: 5-of-73   8.2 GB: 6-of-73   8.2 GB: 7-of-73   8.2 GB: 8-of-73   8.2 GB: 9-of-73   8.2 GB: 10-of-73   8.2 GB: 11-of-73   8.2 GB: 12-of-73   8.2 GB: 13-of-73   8.2 GB: 14-of-73   8.2 GB: 15-of-73   8.2 GB: 16-of-73   8.2 GB: 17-of-73   8.2 GB: 18-of-73   8.2 GB: 19-of-73   8.2 GB: 20-of-73   8.2 GB: 21-of-73   8.2 GB: 22-of-73   8.2 GB: 23-of-73   8.2 GB: 24-of-73   8.2 GB: 25-of-73   8.2 GB: 26-of-73   8.2 GB: 27-of-73   8.2 GB: 28-of-73   8.2 GB: 29-of-73   8.2 GB: 30-of-73   8.2 GB: 31-of-73   8.2 GB: 32-of-73   8.2 GB: 33-of-73   8.2 GB: 34-of-73   8.2 GB: 35-of-73   8.2 GB: 36-of-73   8.2 GB: 37-of-73   8.2 GB: 38-of-73   8.2 GB: 39-of-73   8.2 GB: 40-of-73   8.2 GB: 41-of-73   8.2 GB: 42-of-73   8.2 GB: 43-of-73   8.2 GB: 44-of-73   8.2 GB: 45-of-73   8.2 GB: 46-of-73
Supported Languagesen
Quantization Type4bit
Model ArchitectureInklingForConditionalGeneration
Licenseapache-2.0