What are the hardware requirements for H2ogpt Oasst1 512 12B?

H2ogpt Oasst1 512 12B requires approximately 23.9 GB of VRAM and supports a context window of 2K tokens. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed H2ogpt Oasst1 512 12B and how large is it?

H2ogpt Oasst1 512 12B is developed by h2oai, a model with 12b parameters. The model is published as open weights on Hugging Face and indexed on LLM Explorer with full benchmark history.

How does H2ogpt Oasst1 512 12B perform on standard benchmarks?

H2ogpt Oasst1 512 12B has the following published scores: MMLU 26.01. Compare against reference models on this page or on the LLM Explorer leaderboards.

H2ogpt Oasst1 512 12B by h2oai — VRAM 23.9GB, 2K context

Name: H2ogpt Oasst1 512 12B
Author: h2oai

H2ogpt Oasst1 512 12B is an open-source language model by h2oai. Features: 12b LLM, VRAM: 23.9GB, Context: 2K, License: apache-2.0, LLM Explorer Score: 0.1, Arc: 42.3, HellaSwag: 70.2, MMLU: 26, GSM8K: 1.7.

Dataset:h2oai/openassistant oa... En Gpt Gpt neox Open-source Pytorch Region:us Sharded

Model Card on HF 🤗: https://huggingface.co/h2oai/h2ogpt-oasst1-512-12b

H2ogpt Oasst1 512 12B Parameters and Internals

Model Type

large language model, instruction-following

Supported Languages

languages_supported (en), proficiency_level (unknown)

Training Details

Data Sources:

h2oai/openassistant_oasst1_h2ogpt_graded

Methodology:

Using the fine-tuning dataset

Context Length:

2048

Model Architecture:

GPTNeoXForCausalLM

Input Output

Input Format:

text

Accepted Modalities:

text

Output Format:

text

Performance Tips:

To use effectively, adhere to setup guidelines such as using specific software versions identified in the model card.

LLM Name	H2ogpt Oasst1 512 12B
Repository 🤗	https://huggingface.co/h2oai/h2ogpt-oasst1-512-12b
Model Size	12b
Required VRAM	23.9 GB
Updated	2026-07-14
Maintainer	h2oai
Model Type	gpt_neox
Model Files	5.0 GB: 1-of-5 4.8 GB: 2-of-5 4.9 GB: 3-of-5 5.0 GB: 4-of-5 4.2 GB: 5-of-5
Supported Languages	en
Model Architecture	GPTNeoXForCausalLM
License	apache-2.0
Context Length	2048
Model Max Length	2048
Transformers Version	4.28.1
Tokenizer Class	GPTNeoXTokenizer
Vocabulary Size	50688
Torch Data Type	float16

Rank the H2ogpt Oasst1 512 12B 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 H2ogpt Oasst1 512 12B

Best Alternatives	Context / RAM	Downloads	Likes
Dolly V2 12B	2K / 23.8 GB	3283	1955
...sst Sft 4 Pythia 12B Epoch 3.5	2K / 23.8 GB	818	368
Pythia 12B	2K / 23.8 GB	147000	145
Oasst Rl 1 Pythia 12B	2K / 23.8 GB	10	4
...st Sft 3 Pythia 12B Epoch 2.35	2K / 23.8 GB	10	0
Pythia 12B Deduped	2K / 23.8 GB	4172	53
Oasst Sft 1 Pythia 12B	2K / 23.8 GB	301	277
Pythia 12B Sft V8 7K Steps	2K / 23.8 GB	1106	22
Oasst Pythia 12B Reference	2K / 23.8 GB	1442	0
Pythia 12B Pre V8.12.5K Steps	2K / 23.8 GB	76	6

Note: green Score (e.g. "73.2") means that the model is better than h2oai/h2ogpt-oasst1-512-12b.