LLM EXPLORER 60,828 MODELS INDEXED

MiniMax M2.7 HLWQ Q5 by caiovicentino1

By caiovicentino1 · 17 downloads
HF Repo ↗ Copy URL

MiniMax M2.7 HLWQ Q5 is an open-source language model by caiovicentino1. Features: LLM, VRAM: 42GB, Context: 192K, License: mit, Quantized, LLM Explorer Score: 0.37, ELO: 1404.

  Arxiv:2502.02617   Arxiv:2603.29078 Base model:minimaxai/minimax-m... Base model:quantized:minimaxai...   Conversational   Custom code   En   Hadamard-lloyd-quantization   Hlwq   Long-context   Minimax   Minimax m2   Moe   Polarengine   Q5   Quantized   Region:us   Sharded   Tensorflow
Model Card on HF 🤗: https://huggingface.co/caiovicentino1/MiniMax-M2.7-HLWQ-Q5 

MiniMax M2.7 HLWQ Q5 Benchmarks

LMSys ELO 1404 vs 1272 (so35) 10.4%
nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").

MiniMax M2.7 HLWQ Q5 Parameters and Internals

LLM NameMiniMax M2.7 HLWQ Q5
Repository 🤗https://huggingface.co/caiovicentino1/MiniMax-M2.7-HLWQ-Q5 
Base Model(s)  MiniMaxAI/MiniMax-M2.7   MiniMaxAI/MiniMax-M2.7
Required VRAM42 GB
Updated2026-04-13
Maintainercaiovicentino1
Model Typeminimax_m2
Model Files  2.8 GB: 0-of-130   0.8 GB: 1-of-130   1.6 GB: 2-of-130   0.8 GB: 3-of-130   1.6 GB: 4-of-130   0.8 GB: 5-of-130   1.6 GB: 6-of-130   0.8 GB: 7-of-130   1.6 GB: 8-of-130   0.8 GB: 9-of-130   1.6 GB: 10-of-130   0.8 GB: 11-of-130   1.6 GB: 12-of-130   0.8 GB: 13-of-130   1.6 GB: 14-of-130   0.8 GB: 15-of-130   1.6 GB: 16-of-130   0.8 GB: 17-of-130   1.6 GB: 18-of-130   0.8 GB: 19-of-130   1.6 GB: 20-of-130   0.8 GB: 21-of-130   1.6 GB: 22-of-130   0.8 GB: 23-of-130   1.6 GB: 24-of-130   0.8 GB: 25-of-130   1.6 GB: 26-of-130   0.8 GB: 27-of-130   1.6 GB: 28-of-130   0.8 GB: 29-of-130   1.6 GB: 30-of-130   0.8 GB: 31-of-130   1.6 GB: 32-of-130   0.8 GB: 33-of-130
Supported Languagesen
Quantization Typeq5
Model ArchitectureMiniMaxM2ForCausalLM
Licensemit
Context Length196608
Model Max Length196608
Transformers Version4.46.1
Vocabulary Size200064