UltraChat 200k
Hugging FaceHugging Face H4's filtered 200,000-dialogue slice of the ChatGPT-distilled UltraChat corpus, prepared for supervised fine-tuning of chat models. It anchored the Zephyr recipe and is also used by TinyLlama-Chat, and remains a common starting point for open chat alignment work.
Openness
5 high confidence- license
- MIT
- gated
- false
- dataset_card
- present
MIT-licensed, ungated, full dataset card.
- https://huggingface.co/api/datasets/HuggingFaceH4/ultrachat_200k recorded 2026-09-29
Hub API record for the dataset, whose description opens "Dataset Card for UltraChat 200k", lists the tag "license:mit" and reports it as public and not gated.
- https://huggingface.co/datasets/HuggingFaceH4/ultrachat_200k recorded 2026-08-13
MIT license, ungated access, dataset card
- https://huggingface.co/datasets/HuggingFaceH4/ultrachat_200k/raw/main/README.md recorded 2026-09-29
The raw dataset card declares "license: mit" in its front matter, describes the dataset as "a heavily filtered version of the" UltraChat dataset with a structure section, and has no gating notice.
Adoption
3 high confidenceAdoption is measured by Hugging Face downloads of the dataset over the trailing month, the only usage channel on record for UltraChat 200k.
- https://huggingface.co/api/datasets/HuggingFaceH4/ultrachat_200k recorded 2026-08-13
75,198 downloads in the trailing 30 days for HuggingFaceH4/ultrachat_200k
Capability
3 high confidenceClean SFT before/after in the Zephyr paper (MT-Bench ~6.64), but dropped from 2024-25 default SFT mixes (absent from SmolTalk and Tulu 3).
- https://ar5iv.labs.arxiv.org/html/2310.16944 recorded 2026-09-29
The Zephyr paper states "We first use dSFT, based on the UltraChat" dataset, and its ablation table lists dSFT-1 at "6.64" MT-Bench.
- https://huggingface.co/datasets/allenai/tulu-3-sft-mixture recorded 2026-09-29
The Tulu 3 SFT mixture card says "It contains 939,344 samples from the following sets" and lists FLAN v2, No Robots, "WildChat GPT-4" and others. Searched the body for UltraChat and ultrachat: none.
- https://huggingface.co/datasets/HuggingFaceTB/smoltalk/raw/main/README.md recorded 2026-09-29
The SmolTalk card says "we enhanced our SFT mix by incorporating public datasets to strengthen specific capabilities" such as math and coding. Searched the body for UltraChat and ultrachat: none.
Verified 2026-09-27