LLM EXPLORER 58,873 MODELS INDEXED

Trinity Large Preview 8bit by mlx-community

By mlx-community · 18 downloads
HF Repo ↗ Copy URL

Trinity Large Preview 8bit is an open-source language model by mlx-community. Features: 398.6b LLM, VRAM: 226.6GB, Context: 256K, License: apache-2.0, Quantized, LLM Explorer Score: 0.36, ELO: 1379.

  8-bit   8bit   Afmoe   Ar Base model:arcee-ai/trinity-la... Base model:quantized:arcee-ai/...   Conversational   Custom code   De   En   Es   Fr   Hi   It   Ko   Mlx   Pt   Quantized   Region:us   Ru   Safetensors   Sharded   Tensorflow   Zh
Model Card on HF 🤗: https://huggingface.co/mlx-community/Trinity-Large-Preview-8bit 

Trinity Large Preview 8bit Benchmarks

LMSys ELO 1379 vs 1272 (so35) 8.4%
nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").

Trinity Large Preview 8bit Parameters and Internals

LLM NameTrinity Large Preview 8bit
Repository 🤗https://huggingface.co/mlx-community/Trinity-Large-Preview-8bit 
Base Model(s)  Trinity Large Preview   arcee-ai/Trinity-Large-Preview
Model Size398.6b
Required VRAM226.6 GB
Updated2026-08-10
Maintainermlx-community
Model Typeafmoe
Model Files  4.4 GB: 1-of-82   5.2 GB: 2-of-82   5.1 GB: 3-of-82   5.2 GB: 4-of-82   5.2 GB: 5-of-82   5.1 GB: 6-of-82   5.2 GB: 7-of-82   5.2 GB: 8-of-82   5.1 GB: 9-of-82   5.2 GB: 10-of-82   5.2 GB: 11-of-82   5.1 GB: 12-of-82   5.2 GB: 13-of-82   5.2 GB: 14-of-82   5.1 GB: 15-of-82   5.2 GB: 16-of-82   5.2 GB: 17-of-82   5.1 GB: 18-of-82   5.2 GB: 19-of-82   5.2 GB: 20-of-82   5.1 GB: 21-of-82   5.2 GB: 22-of-82   5.2 GB: 23-of-82   5.1 GB: 24-of-82   5.2 GB: 25-of-82   5.2 GB: 26-of-82   5.1 GB: 27-of-82   5.2 GB: 28-of-82   5.2 GB: 29-of-82   5.1 GB: 30-of-82   5.2 GB: 31-of-82   5.2 GB: 32-of-82   5.1 GB: 33-of-82   5.2 GB: 34-of-82   5.2 GB: 35-of-82   5.1 GB: 36-of-82   5.2 GB: 37-of-82   5.2 GB: 38-of-82   5.1 GB: 39-of-82   5.2 GB: 40-of-82   5.2 GB: 41-of-82   5.1 GB: 42-of-82   5.2 GB: 43-of-82   5.2 GB: 44-of-82
Supported Languagesen es fr de it pt ru ar hi ko zh
Quantization Type8bit
Model ArchitectureAfmoeForCausalLM
Licenseapache-2.0
Context Length262144
Model Max Length262144
Transformers Version4.57.1
Vocabulary Size200192

Best Alternatives to Trinity Large Preview 8bit

Best Alternatives
Context / RAM
Downloads
Likes
Trinity Large Thinking 8bit256K / 221.4 GB271
Trinity Large Thinking256K / 798.4 GB8756186
Trinity Large Preview256K / 798.4 GB313181
Trinity Large Base256K / 798.4 GB37057
Trinity Large TrueBase8K / 798.4 GB28668
Note: green Score (e.g. "73.2") means that the model is better than mlx-community/Trinity-Large-Preview-8bit.