What are the hardware requirements for Ring 2.5 1T?

Ring 2.5 1T requires approximately 273 GB of VRAM and supports a context window of 128K tokens. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed Ring 2.5 1T and how large is it?

Ring 2.5 1T is developed by inclusionAI, a model with 1012.5b parameters. The model is published as open weights on Hugging Face and indexed on LLM Explorer with full benchmark history.

Where can I download or evaluate Ring 2.5 1T?

Ring 2.5 1T is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

Ring 2.5 1T by inclusionAI — VRAM 273GB, 128K context

Name: Ring 2.5 1T
Author: inclusionAI

Ring 2.5 1T is an open-source language model by inclusionAI. Features: 1012.5b LLM, VRAM: 273GB, Context: 128K, License: mit, LLM Explorer Score: 0.31.

Bailing hybrid Compressed-tensors Conversational Custom code Region:us Safetensors Sharded Tensorflow

Model Card on HF 🤗: https://huggingface.co/inclusionAI/Ring-2.5-1T

Ring 2.5 1T Parameters and Internals

LLM Name	Ring 2.5 1T
Repository 🤗	https://huggingface.co/inclusionAI/Ring-2.5-1T
Model Size	1012.5b
Required VRAM	273 GB
Updated	2026-08-08
Maintainer	inclusionAI
Model Type	bailing_hybrid
Model Files	9.7 GB: 0-of-160 3.2 GB: 1-of-160 3.2 GB: 2-of-160 3.2 GB: 3-of-160 3.2 GB: 4-of-160 3.2 GB: 5-of-160 3.2 GB: 6-of-160 3.2 GB: 7-of-160 7.8 GB: 8-of-160 6.5 GB: 9-of-160 6.5 GB: 10-of-160 6.5 GB: 11-of-160 6.5 GB: 12-of-160 6.5 GB: 13-of-160 6.5 GB: 14-of-160 6.5 GB: 15-of-160 8.0 GB: 16-of-160 6.5 GB: 17-of-160 6.5 GB: 18-of-160 6.5 GB: 19-of-160 6.5 GB: 20-of-160 6.5 GB: 21-of-160 6.5 GB: 22-of-160 6.5 GB: 23-of-160 7.8 GB: 24-of-160 6.5 GB: 25-of-160 6.5 GB: 26-of-160 6.5 GB: 27-of-160 6.5 GB: 28-of-160 6.5 GB: 29-of-160 6.5 GB: 30-of-160 6.5 GB: 31-of-160 8.0 GB: 32-of-160 6.5 GB: 33-of-160 6.5 GB: 34-of-160 6.5 GB: 35-of-160 6.5 GB: 36-of-160 6.5 GB: 37-of-160 6.5 GB: 38-of-160 6.5 GB: 39-of-160 7.8 GB: 40-of-160 6.5 GB: 41-of-160 6.5 GB: 42-of-160 6.5 GB: 43-of-160
Model Architecture	BailingMoeV2_5ForCausalLM
License	mit
Context Length	131072
Model Max Length	131072
Transformers Version	4.56.2
Vocabulary Size	157184
Torch Data Type	bfloat16

Rank the Ring 2.5 1T 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