Hugging Face Changelog

The tensor viewer now sends a single request for the headers of all shards instead of fetching each shard one by one, cutting load times on large sharded models such as DeepSeek-V3.2 (9.2s to 0.45s), GLM-5.3 (4.0s to 0.57s), and Qwen2.5-72B-Instruct (1.2s to 0.27s).

Fine-grained access tokens and service account tokens can now be granted a contributor permission that allows creating repositories in an account or selected organizations, with full write access limited to repositories the token created. It mirrors the contribute-repos scope for OAuth apps and targets CI/CD pipelines, training jobs, and agents publishing checkpoints or datasets.

Decision models now have a dedicated decision-model tag, an icon on model pages, and a filter in the Misc section of the Models page. GGUF decision models are tagged automatically from their llama.cpp metadata.

Users can now add a P(doom) estimate, their probability of AI causing an existential catastrophe, to their Hugging Face profile. Answers also feed a community survey on AI risk perception, with only aggregated, anonymized results published in two weeks.

The job page has a new Resources panel next to the logs that shows live CPU, RAM, GPU, and VRAM usage while a job runs, updated every couple of seconds.

Organizations can now subscribe to Hugging Face through their Google Cloud account, putting Inference Providers, Inference Endpoints, Spaces, Jobs, storage, and Team or Enterprise seats on their existing cloud invoice. Subscription works like the AWS Marketplace integration: subscribe from the listing, pick the organization to link, and usage moves to the cloud bill once the subscription is validated.

LeRobot datasets now open with an episode preview on the dataset page, with every camera playing on one shared transport so scrubbing shows each angle at the same moment and hovering the episode list previews each take. Dataset listings also show robot type, episode count, and a first-episode preview.

Feature access can now be controlled per resource group instead of only by organization role, so Jobs can stay open to everyone while Inference Endpoints are restricted to admins and blog publishing is limited to the blog-writer resource group.

Enterprise organizations can now create service accounts: dedicated, organization-owned identities for programmatic access such as CI/CD pipelines, automation scripts, and backend integrations. They aren't tied to any individual, so automated workflows keep running even as…

You can now publish to Hugging Face repositories from GitHub, GitLab, or custom CI systems without adding secrets, using workflow identity federation. This mirrors the secret-less Trusted Publishing experience available for npm and PyPI. CI jobs can also read from gated…

Every Gradio Space now auto-serves an /agents.md endpoint, a machine-readable API description that AI agents can read and call directly. Point your coding agents (like Claude Code, Codex, or Pi) at it and they figure out how to use the Space without any setup.