What are the hardware requirements for Trinity Large Preview?

Trinity Large Preview requires approximately 798.4 GB of VRAM and supports a context window of 256K tokens. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed Trinity Large Preview and how large is it?

Trinity Large Preview is developed by arcee-ai, a model with 398.6b parameters. The model is published as open weights on Hugging Face and indexed on LLM Explorer with full benchmark history.

How does Trinity Large Preview perform on standard benchmarks?

Trinity Large Preview has the following published scores: LMSYS ELO 1379. Compare against reference models on this page or on the LLM Explorer leaderboards.

Trinity Large Preview by arcee-ai — VRAM 798.4GB, 256K context

Name: Trinity Large Preview
Author: arcee-ai

Trinity Large Preview is an open-source language model by arcee-ai. Features: 398.6b LLM, VRAM: 798.4GB, Context: 256K, License: other, LLM Explorer Score: 0.39, ELO: 1379.

Arxiv:2602.17004 Afmoe Ar Base model:arcee-ai/trinity-la... Base model:finetune:arcee-ai/t... Conversational Custom code De En Endpoints compatible Es Eval-results Fr Hi It Ko Pt Region:us Ru Safetensors Sharded Tensorflow Zh

Model Card on HF 🤗: https://huggingface.co/arcee-ai/Trinity-Large-Preview

Trinity Large Preview 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").

What is the LLM Explorer Rank (Score)

Trinity Large Preview Parameters and Internals

LLM Name	Trinity Large Preview
Repository 🤗	https://huggingface.co/arcee-ai/Trinity-Large-Preview
Base Model(s)	Trinity Large Base arcee-ai/Trinity-Large-Base
Model Size	398.6b
Required VRAM	798.4 GB
Updated	2026-07-23
Maintainer	arcee-ai
Model Type	afmoe
Model Files	2.5 GB: 1-of-31 0.7 GB: 2-of-31 0.7 GB: 3-of-31 0.7 GB: 4-of-31 29.4 GB: 5-of-31 29.4 GB: 6-of-31 29.4 GB: 7-of-31 29.4 GB: 8-of-31 29.4 GB: 9-of-31 29.4 GB: 10-of-31 29.4 GB: 11-of-31 29.4 GB: 12-of-31 29.4 GB: 13-of-31 29.4 GB: 14-of-31 29.4 GB: 15-of-31 29.4 GB: 16-of-31 29.4 GB: 17-of-31 29.4 GB: 18-of-31 29.4 GB: 19-of-31 29.4 GB: 20-of-31 29.4 GB: 21-of-31 29.4 GB: 22-of-31 29.4 GB: 23-of-31 29.4 GB: 24-of-31 29.4 GB: 25-of-31 29.4 GB: 26-of-31 29.4 GB: 27-of-31 29.4 GB: 28-of-31 29.4 GB: 29-of-31 29.4 GB: 30-of-31 29.4 GB: 31-of-31
Supported Languages	en es fr de it pt ru ar hi ko zh
Model Architecture	AfmoeForCausalLM
License	other
Context Length	262144
Model Max Length	262144
Transformers Version	4.57.1
Tokenizer Class	PreTrainedTokenizerFast
Padding Token	<\|pad\|>
Vocabulary Size	200192

Rank the Trinity Large Preview capabilities

Have you tried this model? Rate its performance — this feedback helps the ML community find the right model for their needs.

Instruction Following and Task Automation
Factuality and Completeness of Knowledge
Censorship and Alignment
Data Analysis and Insight Generation
Text Generation
Text Summarization and Feature Extraction
Code Generation
Multi-Language Support and Translation

Quantized Models of the Trinity Large Preview

Model	Likes	Downloads	VRAM
Trinity Large Preview 8bit	1	18	226 GB

Best Alternatives to Trinity Large Preview

Best Alternatives	Context / RAM	Downloads	Likes
Trinity Large Thinking	256K / 798.4 GB	8756	186
Trinity Large Base	256K / 798.4 GB	370	57
Trinity Large TrueBase	8K / 798.4 GB	286	68
Trinity Large Thinking 8bit	256K / 221.4 GB	27	1
Trinity Large Preview 8bit	256K / 226.6 GB	18	1

Note: green Score (e.g. "73.2") means that the model is better than arcee-ai/Trinity-Large-Preview.