What are the hardware requirements for InternVL2 5 26B MPO?

InternVL2 5 26B MPO requires approximately 51.2 GB of VRAM. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed InternVL2 5 26B MPO and how large is it?

InternVL2 5 26B MPO is developed by OpenGVLab, a model with 26b 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 InternVL2 5 26B MPO?

InternVL2 5 26B MPO is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

InternVL2 5 26B MPO by OpenGVLab — VRAM 51.2GB

Name: InternVL2 5 26B MPO
Author: OpenGVLab

InternVL2 5 26B MPO is an open-source language model by OpenGVLab. Features: 26b LLM, VRAM: 51.2GB, License: mit, LLM Explorer Score: 0.16.

Arxiv:2312.14238 Arxiv:2404.16821 Arxiv:2411.10442 Arxiv:2412.05271 Base model:finetune:opengvlab/... Base model:opengvlab/internvl2... Conversational Custom code Dataset:opengvlab/mmpr-v1.1 Feature-extraction Image-text-to-text Internvl Internvl chat Multilingual Region:us Safetensors Sharded Tensorboard Tensorflow

Model Card on HF 🤗: https://huggingface.co/OpenGVLab/InternVL2_5-26B-MPO

InternVL2 5 26B MPO Parameters and Internals

LLM Name	InternVL2 5 26B MPO
Repository 🤗	https://huggingface.co/OpenGVLab/InternVL2_5-26B-MPO
Base Model(s)	InternVL2 5 26B OpenGVLab/InternVL2_5-26B
Model Size	26b
Required VRAM	51.2 GB
Updated	2026-08-08
Maintainer	OpenGVLab
Model Type	internvl_chat
Model Files	5.0 GB: 1-of-11 4.9 GB: 2-of-11 4.8 GB: 3-of-11 4.9 GB: 4-of-11 4.9 GB: 5-of-11 5.0 GB: 6-of-11 5.0 GB: 7-of-11 4.9 GB: 8-of-11 5.0 GB: 9-of-11 5.0 GB: 10-of-11 1.8 GB: 11-of-11
Model Architecture	InternVLChatModel
License	mit
Model Max Length	8192
Tokenizer Class	InternLM2Tokenizer
Padding Token	</s>
Torch Data Type	bfloat16

Rank the InternVL2 5 26B MPO 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 InternVL2 5 26B MPO

Model	Likes	Downloads	VRAM
InternVL2 5 26B MPO AWQ	4	128	23 GB

Best Alternatives to InternVL2 5 26B MPO

Best Alternatives	Context / RAM	Downloads	Likes
InternVL2 5 26B AWQ	0K / 23.6 GB	1691	10
InternVL2 5 26B MPO AWQ	0K / 23.6 GB	128	4
InternVL2 26B AWQ	0K / 23.6 GB	76	22

Note: green Score (e.g. "73.2") means that the model is better than OpenGVLab/InternVL2_5-26B-MPO.