LLM EXPLORER 62,323 MODELS INDEXED

MiniMax M2.7 DFlash by nvidia

By nvidia · 61 downloads
HF Repo ↗ Copy URL

MiniMax M2.7 DFlash is an open-source language model by nvidia. Features: 1.3b LLM, VRAM: 2.6GB, Context: 192K, License: other, LLM Explorer Score: 0.38, ELO: 1360.

  Arxiv:2602.06036 Base model:finetune:minimaxai/... Base model:minimaxai/minimax-m...   Dflash   Minimax-m2.7   Model optimizer   Modelopt   Nvidia   Qwen3   Region:us   Safetensors
Model Card on HF 🤗: https://huggingface.co/nvidia/MiniMax-M2.7-DFlash 

MiniMax M2.7 DFlash Benchmarks

LMSys ELO 1360 vs 1272 (so35) 6.9%
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 DFlash Parameters and Internals

LLM NameMiniMax M2.7 DFlash
Repository 🤗https://huggingface.co/nvidia/MiniMax-M2.7-DFlash 
Base Model(s)  MiniMaxAI/MiniMax-M2.7   MiniMaxAI/MiniMax-M2.7
Model Size1.3b
Required VRAM2.6 GB
Updated2026-07-05
Maintainernvidia
Model Typeqwen3
Model Files  2.6 GB
Model ArchitectureDFlashDraftModel
Licenseother
Context Length196608
Model Max Length196608
Vocabulary Size200055
Torch Data Typebfloat16