European Open Source AI Index
DatabaseNewsGuidesAboutContribute

Airoboros

by Jon Durbin

Experimental model tuned primarily from synthetic data generated by the airoboros implementation of LLM self-instruction.
Text
Full
https://huggingface.co/jondurbin/airoboros-dpo-110b-3.3
Qwen1.5-110B
Airoboros-DPO-110B-3.3
Purposely left ambiguous
Individual. Primary contributor to chutes.ai.
https://github.com/jondurbin
May 2024
Availability
Base Model Data
Base model data sources nowhere documented or specified.
End User Model Data
Most data generated by airoboros, an implementation of the Self-Instruct paper. Many other data sources linked. One data source links to a 404 page. Additional data sources linked in tags but not outlined in model card.
https://huggingface.co/datasets/jondurbin/airoboros-3.2https://huggingface.co/datasets/bluemoon-fandom-1-1-rp-cleanedhttps://huggingface.co/datasets/boolqhttps://huggingface.co/datasets/jondurbin/gutenberg-dpo-v0.1https://huggingface.co/datasets/LDJnr/Capybarahttps://huggingface.co/datasets/jondurbin/cinematika-v0.1https://huggingface.co/datasets/glaiveai/glaive-function-calling-v2https://huggingface.co/datasets/grimulkan/LimaRP-augmentedhttps://huggingface.co/datasets/piqahttps://huggingface.co/datasets/Vezora/Tested-22k-Python-Alpacahttps://huggingface.co/datasets/mattpscott/airoboros-summarizationhttps://huggingface.co/datasets/unalignment/toxic-dpo-v0.2
Base Model Weights
Weights available through HuggingFace.
https://huggingface.co/Qwen/Qwen1.5-110B
End User Model Weights
Weights available through HuggingFace.
https://huggingface.co/jondurbin/airoboros-dpo-110b-3.3
Training Code
Repo exists only for general fine-tuning implementation. Earlier airoboros model versions included some architecture and training information. Latest versions (>3.1.2) do not.
https://github.com/jondurbin/qlora
Documentation
Code Documentation
Code is not very systematically documented.
https://github.com/jondurbin/qlora
Hardware Architecture
Earlier airoboros model versions included some architecture and training information. Latest versions (>3.1.2) do not.
Preprint
No preprint found.
Paper
No peer-reviewed paper found.
Modelcard
Model card contains some information, mainly relating to inference and licensing.
https://huggingface.co/jondurbin/airoboros-dpo-110b-3.3
Datasheet
Datasheets contain differing levels of documentation. Exact data used as well as data collection and curation procedure unknown. No data sheet for base model data exists.
https://huggingface.co/datasets/jondurbin/airoboros-3.2https://huggingface.co/datasets/bluemoon-fandom-1-1-rp-cleanedhttps://huggingface.co/datasets/boolqhttps://huggingface.co/datasets/jondurbin/gutenberg-dpo-v0.1https://huggingface.co/datasets/LDJnr/Capybarahttps://huggingface.co/datasets/jondurbin/cinematika-v0.1https://huggingface.co/datasets/glaiveai/glaive-function-calling-v2https://huggingface.co/datasets/grimulkan/LimaRP-augmentedhttps://huggingface.co/datasets/piqahttps://huggingface.co/datasets/Vezora/Tested-22k-Python-Alpacahttps://huggingface.co/datasets/mattpscott/airoboros-summarizationhttps://huggingface.co/datasets/unalignment/toxic-dpo-v0.2
Access
Package
Pip package available, however only supports older versions of airoboros.
https://pypi.org/project/airoboros/
API and Meta Prompts
No API found.
Licenses
Licensing left ambiguous because of murky status of OpenAI-derived Self-Instruct data.
https://huggingface.co/jondurbin/airoboros-dpo-110b-3.3#licence-and-usage-restrictions
Last updated 09 June 2025
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 10 March 2026, website content last updated 11 March 2026.