---
title: "State of Open Models: Summer 2026 Observations"
description: "The report analyzes Hugging Face Hub activity from January through August 2026 to assess how open-model releases, usage, licensing, and agent traffic are changing. Public repositories expanded from 2.43 to 2.96 million models, while 85.6% of models had fewer than 200 lifetime downloads and 1.5% of repositories captured 99.2% of downloads. Chinese labs reached monthly ceilings of 754B to 2.78T parameters, while broader portfolios such as Qwen’s generated far more downloads than frontier-only strategies; likes and downloads measured attention and infrastructure dependence differently. Large Chinese releases were often Apache 2.0 or MIT licensed, though newer trillion-parameter models introduced restrictions, while agents became the Hub’s leading user and autonomous-agent activity exposed the practical value of running quantized open models locally."
---

# State of Open Models: Summer 2026 Observations

[Hugging Face](https://yomu.fyi/company/hugging-face) · Adina Yakefu, Apolinário from multimodal AI art, Irene Solaiman · Aug 14, 2026

**Type:** Benchmark

## Summary

The report analyzes Hugging Face Hub activity from January through August 2026 to assess how open-model releases, usage, licensing, and agent traffic are changing. Public repositories expanded from 2.43 to 2.96 million models, while 85.6% of models had fewer than 200 lifetime downloads and 1.5% of repositories captured 99.2% of downloads. Chinese labs reached monthly ceilings of 754B to 2.78T parameters, while broader portfolios such as Qwen’s generated far more downloads than frontier-only strategies; likes and downloads measured attention and infrastructure dependence differently. Large Chinese releases were often Apache 2.0 or MIT licensed, though newer trillion-parameter models introduced restrictions, while agents became the Hub’s leading user and autonomous-agent activity exposed the practical value of running quantized open models locally.

## Context

The report presents observations from Hugging Face Hub activity during the first seven months of 2026, with an edit covering early August. It examines model and dataset growth, parameter-scale competition between Chinese and U.S. labs, the relationship between attention and usage, licensing, ecosystem adoption, and agent traffic. The authors caution that Hub metrics do not directly measure model quality, commercial adoption, or total market share.

## Approach / What changed

The analysis compares repository counts, downloads, likes, derivatives, model releases, parameter sizes, licensing terms, publisher portfolios, and agent-tagged traffic. It uses download activity within the reporting window, controls comparisons for repository age where stated, and separates likes as a measure of attention from downloads as a measure of recurring pipeline usage. It also documents changes to machine-readable Hub interfaces and reports a technical timeline of an autonomous-agent intrusion.

## Takeaways

- Public model repositories grew from 2.43 million to 2.96 million, but usage remained highly concentrated: 85.6% of models had fewer than 200 lifetime downloads, while 1.5% of repositories accounted for 99.2% of downloads.
- Likes and downloads reflect different behaviors. No model published in 2026 entered the year-to-date download top 25, while thirteen entries dated from 2022; all-MiniLM-L6-v2 reached 1.55 billion downloads against 5,156 likes.
- Chinese labs released many large models under permissive licenses, including MIT and Apache 2.0, but Kimi K3 and Qwen 3.8 2.4T introduced non-commercial restrictions and revenue-share requirements. Agents became the Hub’s leading user category by the report’s conclusion.

**Tags:** [Developer Experience](https://yomu.fyi/topic/developer-experience), [LLMs](https://yomu.fyi/topic/llm), [Machine Learning](https://yomu.fyi/topic/machine-learning), [Open Source](https://yomu.fyi/topic/open-source)

- Source: [Hugging Face](https://huggingface.co/blog/state-of-open-models-summer-2026)
- Source URL: https://huggingface.co/blog/state-of-open-models-summer-2026
- Ingested by Yomu: 2026-08-29T05:41:24.389Z

[Read original post](https://huggingface.co/blog/state-of-open-models-summer-2026)