European Open Source AI Index

Database News Guides About Contribute

NeuralChat

by Intel

About the model:

Mistral-based Orca-finetuned chat model.

Model type:

Text

Model performance class:

Full

Link to the model:

https://huggingface.co/Intel/neural-chat-7b-v3-3

Base models:

Mistral-7B-v0.1

End model:

Neural-Chat-7B-v3.3

End model license:

Apache-2.0

About the organisation:

Intel, a major chip manufacturer.

Link to the organisation:

https://huggingface.co/Intel

Model release date:

November 2023

Availability

Base Model Data

Mistral has not disclosed anything about its training data.

https://huggingface.co/mistralai/Mistral-7B-v0.1/discussions/8

End User Model Data

Dataset used for fine-tuning is shared and available

https://huggingface.co/datasets/Open-Orca/SlimOrca https://huggingface.co/datasets/meta-math/MetaMathQA https://huggingface.co/datasets/Intel/orca_dpo_pairs

Base Model Weights

Weights available through HuggingFace.

https://huggingface.co/mistralai/Mistral-7B-v0.1

End User Model Weights

Finetuned model made openly available by Intel.

https://huggingface.co/Intel/neural-chat-7b-v3-3

Training Code

Mistral base model is not open, but repo gives sample code for fine-tuning based on that.

https://github.com/intel/intel-extension-for-transformers/tree/main/intel_extension_for_transformers/neural_chat/examples/finetuning/finetune_neuralchat_v3

Documentation

Code Documentation

Mistral remains closed so only documentation pertains to fine-tuning steps, but that code is quite well documented, so partial.

https://github.com/intel/intel-extension-for-transformers/tree/main/intel_extension_for_transformers/neural_chat/examples/finetuning/finetune_neuralchat_v3#how-to-train-intelneural-chat-7b-v3-1-on-intel-gaudi2

Hardware Architecture

Described on HuggingFace model card and a Medium post.

https://medium.com/intel-analytics-software/the-practice-of-supervised-finetuning-and-direct-preference-optimization-on-habana-gaudi2-a1197d8a3cd3

Preprint

A medium post is apparently the only scientific documentation of this model.

https://medium.com/intel-analytics-software/the-practice-of-supervised-finetuning-and-direct-preference-optimization-on-habana-gaudi2-a1197d8a3cd3

Paper

No peer-reviewed paper found.

Modelcard

Model card provides limited information in desired categories.

https://huggingface.co/Intel/neural-chat-7b-v3-3

Datasheet

Datasheets contain differing amounts of information. No datasheet available for the base model.

https://huggingface.co/datasets/Open-Orca/SlimOrca https://huggingface.co/datasets/meta-math/MetaMathQA https://huggingface.co/datasets/Intel/orca_dpo_pairs

Access

Licenses

Apache-2.0, an OSI-approved license.

https://huggingface.co/Intel/neural-chat-7b-v3-1#fp32-inference-with-transformers

Is this information not up to date?

Contribute here ->

Supported by the Centre for Language Studies and the Dutch Research Council. Website design & development © 2024 by BSTN. This version of the index generated 09 April 2026, website content last updated 11 March 2026.