Zephyr 7B β

Available · Language

Zephyr 7B β is the second model in Hugging Face's Zephyr series, a 7B chat assistant fine-tuned from Mistral 7B v0.1 and published as open weights under the MIT licence. Its technical report describes distilled supervised fine-tuning followed by distilled preference optimization on AI feedback, without human annotation; the training code followed in the Alignment Handbook. [1] [2] [3]README, News: November 10, 2023 Primary source

Timeline of Zephyr 7B β →

Claims and evidence

  • Released The technical report was submitted on 2023-10-25 and the public model repository was created on 2023-10-26; no announcement giving a single release day was found, so only the month is recorded. Primary source[2]arXiv submission history, v1 of 2023-10-25 [1]Hub commit history: repository created 2023-10-26
  • Status AvailableWeights are downloadable from the HuggingFaceH4 organisation on the Hub. Hugging Face publishes no deprecation page for its models. Primary source[1]Hub repository HuggingFaceH4/zephyr-7b-beta, Files and versions
  • Change · Training data Trained on a filtered and preprocessed version of UltraChat and a binarized version of UltraFeedback, whose completions were scored by GPT-4.Compared with Zephyr 7B Alpha Primary source[1]
  • Input text Primary source[1]first paragraph; Model description
  • Output text Primary source[1]first paragraph; Model description
  • Open weights YesMIT licence. Primary source[1]Model description, License
  • Parameters 7B Primary source[1]Model description, Model type
  • Access Open-weights download Primary source[1]Intended uses and limitations

Lineage

Predecessors

Successors

Based on

  • Mistral 7B · Mistral AI · 27 September 2023 · derived (fine tune)

Variants and derived

None recorded.

Siblings

None recorded.

All ancestors

All descendants

Variants

No variants recorded in this record.

Related AI Radar coverage

AI Radar coverage starts in June 2026; no coverage linked yet.

All model releases from Hugging Face on AI Radar →