8 models across 3 families · Latest: StarChat2 15B (2024-07)
Community-driven open-source AI model hub
Hugging Face H4's portfolio covers 8 active models across 3 non-obsolete families, with task labels spanning coding, classification, and json / tool use. Open a model detail page to compare provider routes and sourced benchmarks.
Portfolio context: 3 decision-task tags, 8 active tracked models, latest research stamp 2026-05-19.
Use this portfolio page for
- Teams evaluating coding, classification, and json / tool use across this lab's releases
- Readers comparing families before locking a flagship SKU
- 8 tracked SKUs for migration and pricing follow-ups
Do not stop here for
- Choosing a hosting provider without opening a model page for price ladders
Active models
8
Non-deprecated SKUs linked to this researcher
Active families
3
Non-obsolete families in coverage
Open catalog
4 OSS
0 open-weight (text match)
Decision task tags
3
Mapped to the site-wide task taxonomy
Latest dated release
2024-07-04
StarChat2 15B
Freshness
2026-05-19
Researched 16d ago
Release cadence
Showing 5 recent dated ships (full timeline below). Latest spotlight: StarChat2 15B (2024-07-04).
Where this lab wins
- Coding: 1 tracked model with SWE-bench / HumanEval-style scores.
- Classification: 2 tracked models with MMLU-class moderation/safety coverage.
- JSON/tool-use: 1 tracked model with BFCL / Nexus strict-JSON routing coverage.
Flagship quality / price signal
Anchor SKU: Zephyr 7B Beta (best sourced coding Q/$ in this portfolio).
Coding task grade A · benchmark humaneval score 67.8 · cheapest tracked output from ladder $0.200 per 1M tokens
Hugging Face H4 is an American AI research organization founded in 2016. Community-driven open-source AI model hub. Hugging Face H4 ships 3 model families totaling 8 models, with the most recent release StarChat2 15B in 2024-07. Notable families include StarChat2, Zephyr, and StarChat. Use it as a stable reference for lab background, release coverage, and follow-up model pages as they are added. View official API endpoints, benchmark performance, and coding/agent fit for every Hugging Face H4 model.
About
Hugging Face, an influential player in the AI landscape, revolutionized the field through its groundbreaking contributions to generative AI and Large Language Models (LLMs). Established in 2016, the company initially embarked on a journey to create a playful chatbot for teenagers. However, recognizing the broader potential of its technology, Hugging Face pivoted towards developing robust tools for machine learning, focusing on natural language processing (NLP). This strategic change led to the development of the Transformers library, a monumental open-source toolkit that democratized access to state-of-the-art AI capabilities. Its success is attributed to its user-friendly design and detailed documentation, which made cutting-edge AI technology accessible to a wider audience. The open-source nature of Hugging Face has been pivotal in fostering a collaborative and dynamic community of developers and researchers. This community actively contributes to the platform's growth, leading to the creation of an extensive library of pre-trained models and datasets across diverse applications, such as language translation and sentiment analysis. Today, Hugging Face's platform boasts over 100,000 models and 20,000 datasets, an impressive scale that underscores its impact on the AI field. More than 10,000 companies leverage its tools for various AI and machine learning applications, attesting to its broad applicability and influence. Hugging Face's commitment goes beyond technological advancements; it champions ethical AI practices. Through its Model Card initiative, the company aims to enhance transparency by providing detailed documentation on the attributes and potential biases of its models. This commitment to responsible AI development is coupled with initiatives to encourage open discussions on ethical AI issues. Additionally, Hugging Face has formed strategic partnerships with tech giants like Google, Amazon, and Microsoft, integrating its cutting-edge tools into the core operations of leading companies, further cementing its position as a leader in AI innovation. With a valuation of $4.5 billion in 2024, Hugging Face not only reflects impressive financial success but also signifies its profound influence on the future of AI.
Featured models
| Model | Released | Context | Input price ($/1M) | Output price ($/1M) | License |
|---|---|---|---|---|---|
| StarChat2 15B | 2024-07-04 | 16k | - | - | Open Source |
| Zephyr 7B Alpha | 2023-10-26 | - | $0.05 | $0.25 | Unknown |
| Zephyr 7B Beta | 2023-10-26 | - | $0.20 | $0.20 | Unknown |
Model families
Recent releases
- StarChat2 15B- 2024-07-04
- Zephyr 7B Alpha- 2023-10-26
- Zephyr 7B Beta- 2023-10-26
- Zephyr 7B Gemma- 2023-10-26
- Zephyr ORPO 141B- 2023-10-26
FAQ
Who founded Hugging Face H4 and when?
Hugging Face H4 was founded in 2016 and is associated with New York City, New York, United States.
What models has Hugging Face H4 released?
Hugging Face H4 ships 8 models across 3 families: StarChat2, Zephyr, and StarChat.
Is Hugging Face H4's technology open source?
Some tracked Hugging Face H4 models are open-weight, including StarChat2 15B, Fireworks Zephyr-7B-beta, and StarChat Alpha.
Where is Hugging Face H4 headquartered?
Hugging Face H4 is headquartered in New York City, New York, United States.
What is Hugging Face H4 known for?
Community-driven open-source AI model hub. Its most prominent tracked family is StarChat2.
How can I access Hugging Face H4's models?
Hugging Face H4's models are available via Baseten API, DeepInfra, Fireworks AI, and Replicate API.
Explore related pages
Last reviewed: 2026-05-19. Data sourced from public lab announcements and provider documentation.