LLM EXPLORER 58,975 MODELS INDEXED

MiniMax M2.5 Int4 AutoRound by Intel

By Intel · 405 downloads
HF Repo ↗ Copy URL

MiniMax M2.5 Int4 AutoRound is an open-source language model by Intel. Features: 31.7b LLM, VRAM: 121.4GB, Context: 192K, LLM Explorer Score: 0.24.

  Arxiv:2309.05516   4-bit   Auto-round Base model:minimaxai/minimax-m... Base model:quantized:minimaxai...   Conversational   Custom code   Minimax m2   Region:us   Safetensors   Sharded   Tensorflow
Model Card on HF 🤗: https://huggingface.co/Intel/MiniMax-M2.5-int4-AutoRound 

MiniMax M2.5 Int4 AutoRound Parameters and Internals

LLM NameMiniMax M2.5 Int4 AutoRound
Repository 🤗https://huggingface.co/Intel/MiniMax-M2.5-int4-AutoRound 
Base Model(s)  MiniMaxAI/MiniMax-M2.5   MiniMaxAI/MiniMax-M2.5
Model Size31.7b
Required VRAM121.4 GB
Updated2026-08-07
MaintainerIntel
Model Typeminimax_m2
Model Files  5.4 GB: 1-of-23   5.4 GB: 2-of-23   5.4 GB: 3-of-23   5.4 GB: 4-of-23   5.4 GB: 5-of-23   5.4 GB: 6-of-23   5.4 GB: 7-of-23   5.4 GB: 8-of-23   5.4 GB: 9-of-23   5.4 GB: 10-of-23   5.4 GB: 11-of-23   5.4 GB: 12-of-23   5.4 GB: 13-of-23   5.4 GB: 14-of-23   5.4 GB: 15-of-23   5.4 GB: 16-of-23   5.4 GB: 17-of-23   5.4 GB: 18-of-23   5.4 GB: 19-of-23   5.4 GB: 20-of-23   5.4 GB: 21-of-23   5.4 GB: 22-of-23   2.6 GB: 23-of-23
Model ArchitectureMiniMaxM2ForCausalLM
Context Length196608
Model Max Length196608
Transformers Version4.57.1
Tokenizer ClassGPT2Tokenizer
Vocabulary Size200064