LLM EXPLORER 62,155 MODELS INDEXED

MiMo V2.6 Flash RL MLX REAP37 Mixed GPTQ MTP by tacodevs

By tacodevs · 368 downloads
HF Repo ↗ Copy URL

MiMo V2.6 Flash RL MLX REAP37 Mixed GPTQ MTP is an open-source language model by tacodevs. Features: 195.2b LLM, VRAM: 100.6GB, Context: 1024K, License: mit, Quantized, LLM Explorer Score: 0.32.

  Arxiv:2510.13999   Arxiv:2605.23078   8-bit   Apple-silicon Base model:quantized:xiaomimim... Base model:xiaomimimo/mimo-v2....   Conversational   Gptq   Mimo-v2   Mimo v2 flash   Mixture-of-experts   Mlx   Mtp   Quantized   Reap   Region:us   Safetensors   Sharded   Tensorflow
Model Card on HF 🤗: https://huggingface.co/tacodevs/MiMo-V2.6-Flash-RL-MLX-REAP37-mixed-GPTQ-MTP 

MiMo V2.6 Flash RL MLX REAP37 Mixed GPTQ MTP Parameters and Internals

LLM NameMiMo V2.6 Flash RL MLX REAP37 Mixed GPTQ MTP
Repository 🤗https://huggingface.co/tacodevs/MiMo-V2.6-Flash-RL-MLX-REAP37-mixed-GPTQ-MTP 
Base Model(s)  XiaomiMiMo/MiMo-V2.6-Flash-RL   XiaomiMiMo/MiMo-V2.6-Flash-RL
Model Size195.2b
Required VRAM100.6 GB
Updated2026-09-25
Maintainertacodevs
Model Typemimo_v2_flash
Model Files  4.4 GB: 1-of-22   4.5 GB: 2-of-22   5.0 GB: 3-of-22   4.5 GB: 4-of-22   5.0 GB: 5-of-22   4.5 GB: 6-of-22   5.0 GB: 7-of-22   4.9 GB: 8-of-22   4.8 GB: 9-of-22   4.8 GB: 10-of-22   4.5 GB: 11-of-22   4.5 GB: 12-of-22   4.5 GB: 13-of-22   4.5 GB: 14-of-22   4.5 GB: 15-of-22   4.5 GB: 16-of-22   4.5 GB: 17-of-22   4.5 GB: 18-of-22   4.5 GB: 19-of-22   4.5 GB: 20-of-22   4.5 GB: 21-of-22   3.7 GB: 22-of-22
GPTQ QuantizationYes
Quantization Typegptq
Model ArchitectureMiMoV2FlashForCausalLM
Licensemit
Context Length1048576
Model Max Length1048576
Transformers Version5.3.0
Tokenizer ClassQwen2Tokenizer
Padding Token<|endoftext|>
Vocabulary Size152576
Errorsreplace