What are the hardware requirements for DialoGPT Medium?

DialoGPT Medium requires approximately 0.9 GB of VRAM. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed DialoGPT Medium and how large is it?

DialoGPT Medium is developed by microsoft. The model is published as open weights on Hugging Face and indexed on LLM Explorer with full benchmark history.

How does DialoGPT Medium perform on standard benchmarks?

DialoGPT Medium has the following published scores: MMLU 25.84. Compare against reference models on this page or on the LLM Explorer leaderboards.

DialoGPT Medium by microsoft — VRAM 0.9GB

Name: DialoGPT Medium
Author: microsoft

DialoGPT Medium is an open-source language model by microsoft. Features: LLM, VRAM: 0.9GB, License: mit, LLM Explorer Score: 0.14, Arc: 24.5, HellaSwag: 26.2, MMLU: 25.8.

Arxiv:1911.00536 Conversational Endpoints compatible Gpt2 Jax Pytorch Region:us Rust Tf

Model Card on HF 🤗: https://huggingface.co/microsoft/DialoGPT-medium

DialoGPT Medium Parameters and Internals

Model Type

Conversational, Dialogue Response Generation

Use Cases

Areas:

Research, Commercial applications

Applications:

Chatbots, Virtual assistants

Primary Use Cases:

Dialogue generation

Additional Notes

Human evaluation results indicate high response quality.

Supported Languages

English (High proficiency)

Training Details

Data Sources:

Reddit discussion thread

Data Volume:

147M multi-turn dialogues

Methodology:

Pretraining on large-scale dialogue data

Input Output

Input Format:

User input encoded with tokenizer

Accepted Modalities:

Text

Output Format:

Generated dialogue response

LLM Name	DialoGPT Medium
Repository 🤗	https://huggingface.co/microsoft/DialoGPT-medium
Required VRAM	0.9 GB
Updated	2026-07-31
Maintainer	microsoft
Model Type	gpt2
Model Files	0.9 GB
Model Architecture	GPT2LMHeadModel
License	mit
Model Max Length	1024
Tokenizer Class	GPT2Tokenizer
Vocabulary Size	50257
Activation Function	gelu_new
Errors	replace

Rank the DialoGPT Medium 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

Best Alternatives to DialoGPT Medium

Best Alternatives	Context / RAM	Downloads	Likes
Judge GPT2	0.5K / 0.4 GB	10	0
Sanitizer Test Evocati	0K / GB	482	1
Gpt2 Base Bne	0K / 0.5 GB	1736	0
Hanuman	0K / 0.5 GB	17	0
BrtGPT 1 Pre	0K / 0.4 GB	132	0
GPT 2 Large 115K Steps	0K / 0 GB	11	0
GPT 2 Large 51K Steps	0K / 0 GB	7	0
GPT 2 Large 40K Steps	0K / 0 GB	10	0
GPT 2 Large 32K Steps	0K / 0 GB	12	0
GPT 2 Large 20K Steps	0K / 0 GB	9	0

Note: green Score (e.g. "73.2") means that the model is better than microsoft/DialoGPT-medium.