Hugging Face Release notes and new features

Transformers Patch release v5.22.2

This patch release includes security hardening for Jinja sandboxing and fixes for distributed training using NCCL.

Transformers Patch release v5.22.1

Transformers patch release v5.22.1 fixes documentation errors.

[v2.4.0] Copy files into Buckets with `HfFileSystem`, outdated CLI extension hints

huggingface_hub v2.4.0 adds support for copying files into Buckets using HfFileSystem across bucket-to-bucket and repository workflows.

Transformers 5.22.0: Mistral Large 4 (Le Chonk)

Transformers v5.22.0 adds support for Mistral Large 4, a 1-trillion parameter multimodal Mixture-of-Experts model featuring a 1-million token context window.

• Transformers Patch release v5.21.1
• Transformers 5.22.0: Mistral Large 4 (Le Chonk)

Transformers Release: v5.21.0

Transformers v5.21.0 introduces the MiniCPM-OCR document-understanding model and the Flint mixture-of-experts language model. These additions expand the library's support for specialized vision and language tasks.

Diffusers 0.42.0: FLUX.3, Fibo 2.0, Kandinsky 6, and more

Diffusers 0.42.0 introduces support for FLUX.3 Image Klein, Bria Fibo 2.0, and Kandinsky 6, along with new modular pipeline capabilities.

[v2.3.0] Bucket live-follow, faster `HfFileSystem` listings and more

The HfFileSystem now supports live_follow=True, which automatically refreshes cached directory listings by receiving file change notifications from the Hub.

Transformers Patch release v5.20.2

Patch release: v5.20.2 This patch notably solves an issue where some inputs could loose their original type inside prepare_inputs_for_generation in generate, leading to problems with multimodal models. It also fixes some other issues listed on the below commits.

Transformers Release v5.20.1

Release v5.20.1 Special surprise including the new model Beam from Reflection AI :partying_face: Beam is a token-efficient, frontier-scale model focused on coding and agentic use cases. It is built for teams that want to run a powerful model on their own hardware, infrastructure, or cloud, and customize it to their specific needs.

Transformers Release v5.20.0

Release v5.20.0 New Model additions Muse Spark Open Muse Spark Open is a multimodal mixture-of-experts model from Meta that takes text, images, audio, and video as input and generates text. It has 816B total parameters (42B active per token) and a context window of up to 1,048,576 tokens shared across modalities.

[v2.2.0] `hf jobs stats` snapshots, download and `HfFileSystem` fixes

A small release this week, mostly bug fixes and quality-of-life improvements. hf jobs stats now returns right away instead of streaming forever, concurrent downloads reliably use hf_xet, and HfFileSystem caching is more predictable. Two behavior changes are listed under Breaking changes.

Diffusers 0.41.0: QwenImage 2.1 pipeline and more

> [!TIP] > This release brings Qwen-Image 2.1 to Diffusers, with text-to-image generation, image editing, native transparency, and LoRA training. > > Starting with this release, we’re adopting the Transformers release philosophy: coordinating minor releases around new model integrations, while including the other changes merged since the previous release.

Transformers Release v5.19.0

Release v5.19.0 New Model additions EmbeddingGemma2 EmbeddingGemma 2 is a multimodal embedding model from Google built on the Gemma 4 architecture. It encodes text, images, audio, and video, individually or combined in one input, into a shared 768-dimensional vector space for cross-modal retrieval, semantic similarity, clustering, and classification.

[v2.1.0] Better Jobs tooling, ZeroGPU quota tracking, and revamped Model Catalog

📊 Jobs: retries, rescheduling, and port exposure Hugging Face Jobs get a batch of quality-of-life improvements this release. Jobs can now retry automatically on failure with the new attempts parameter, and you can start a fresh Job from an existing Job's saved spec (including secrets and hardware settings) with rerun_job / hf jobs rerun.

• [v2.1.0] Better Jobs tooling, ZeroGPU quota tracking, and revamped Model Catalog
• Hugging Face Hub 2.1.1

Transformers Release 5.18.0

New Model additions Nemotron 3 Diarization Nemotron 3 Diarization is an open-weight streaming speaker diarization model designed to determine "who spoke when" in real-world audio. It supports both streaming and offline inference, handles up to eight speakers, and orders speaker outputs by each speaker's first arrival in the input audio.

[v1.32.0] Shared blob store, sandbox security hardening and faster imports

📂 [Cache] Shared blob store: deduplicate Xet files across repos The cache now deduplicates Xet files across repos. A Xet file downloaded through hf_xet is stored once at <CACHE_DIR>/blobs/<prefix>/<xet_hash> and every repo that needs it gets a relative symlink instead of a download: no bytes are transferred and no extra space is used, even across different...

[v1.31.0] Custom labels for Sandboxes, More resilient downloads and more

🏷️ Custom labels for Sandboxes Dedicated sandboxes now accept custom labels, attached to the underlying Job. This is useful for cost attribution, bookkeeping, or finding and reconnecting to the sandboxes created by a given controller run. Labels use the same -l / --label KEY=VALUE syntax as hf jobs run .

Transformers Release: v5.16.0

Release v5.16.0 New Model additions Qwen4-Exp Qwen4-Exp builds on Qwen3.5's hybrid text and multimodal architecture with three key components: GatedResidual (GR), Qwen Sparse Attention (QSA), and Per-Layer Embedding (PLE). GR is a Qwen-developed residual architecture that combines Hyper-Connection with GatedNorm.

• Transformers Release v5.16.1
• Transformers Release: v5.16.0

Diffusers 0.40.0: New pipelines, tensor-parallel support, improved CLI, and more

Tip This release features several new pipelines, including LTX2.5, MiniMax H3, and Wan Animate 2. We're also graduating Modular Diffusers out of the experimental phase and announcing its stable support. Additionally, this release includes minimal support for tensor-parallel. There's a lot more that went down in this release.

Transformers Patch release: v5.15.1

Patch release v5.15.1 This patch most notably solves a few issues with DFlash and MTP candidate generators, as well as an issue where images could sometimes not be processed on accelerator if using Lanczos filter.

Transformers Release: v5.15.0

Release v5.15.0 New Model additions Meta Muse Glimmer Muse Glimmer, released today, is Meta’s new multimodal model, especially designed for agentic use cases. Distilled from Muse to 30B parameters, and released under the Apache 2.0 license, it can be deployed to local setups for privacy-aware applications such as coding, document analysis, personal...

Sandboxes for the Hugging Face MCP Server

Sandboxes give an assistant secure execution environments attached to buckets and repositories for dataset analysis, model training, and Space creation.

• Sandboxes for the Hugging Face MCP Server
• Hugging Face MCP Server adds hf_fs tool

Egress metrics for users and organizations

Users see egress usage in the dashboard and organizations get a per-user egress breakdown. It currently covers only traffic routed through the Hugging Face CDN.

Build Spaces with AI agents

The new Space creation page has an option to build with an AI agent: copy the generated command into your agent to build and iterate on a Space for a model, paper, or local folder.

Transformers Patch release: v5.14.1

Patch release v5.14.1 This patch solves a few issues which appeared when integrating Inkling model, most notably an issue affecting models using EncoderDecoderCache during assisted generation. It also fixes an issue that could appear during prefill with StaticCache and sdpa without padding for Inkling which uses a position_bias.

Transformers Release v5.14.0

Release v5.14.0 New Model additions Inkling (fresh from Thinking Machines): 975B total, 41B active Add Inkling model #47347 by @molbap @Cyrilvallez @eustlb and @zucchini-nlp Inkling is a general-purpose multimodal model that accepts text, image and audio inputs and generates text outputs.

Fine-grained access token presets

Fine-grained tokens can be created from presets (Read-Only, Inference, Write, CI/CD, Full Access), and preset URLs link straight to a pre-configured token.

Transformers Patch release v5.13.1

Patch release v5.13.1 This patch is focused on enabling transformers for the latest release of vllm! Be more defensive with remap_legacy_layer_types for custom models ( #47245 ) from @hmellor Fix custom code which doesn't know about the new linear layer type names ( #47174 ) from @hmellor Fix case where _LazyAutoMapping.register is passed a str key ( #47148...

Diffusers 0.39.0: New image and video pipelines, core library improvements, and more

New Pipelines Cosmos 3 Cosmos 3 is NVIDIA's unified world foundation model (WFM) for Physical AI , a single omni-model built on a Mixture-of-Transformers (MoT) architecture that combines world generation, physical reasoning, and action generation, replacing the separate Predict, Reason, and Transfer models from earlier Cosmos releases.

Transformers Release v5.13.0

Release v5.13.0 New Model additions KimiK 2.5, 2.6, and 2.7 This release includes the architecture for Kimi 2.5 which is used by 2.5-2.7: Kimi K2.5 is an open-source, native multimodal agentic model that advances practical capabilities in long-horizon coding, coding-driven design, proactive autonomous execution, and swarm-based task orchestration.

Filter Models page by hardware

A Hardware filter on the Models page shows models that fit a specific GPU, CPU, or Apple Silicon chip, set from Hardware settings and shareable via URL.

Send Hub feedback from the user menu

Users can share feedback on the Hub, including bug reports and feature suggestions, directly with the Hugging Face team from the user menu.

Diffusers 0.38.0: New image and audio pipelines, Core library improvements, and more

New Pipelines LLaDA2 LLaDA2 is a family of discrete diffusion language models that generate text through block-wise iterative refinement. Instead of autoregressive token-by-token generation, LLaDA2 starts with a fully masked sequence and progressively unmasks tokens by confidence over multiple refinement steps.

Diffusers 0.37.0: Modular Diffusers, New image and video pipelines, multiple core library improvements, and more 🔥

Modular Diffusers Modular Diffusers introduces a new way to build diffusion pipelines by composing reusable blocks. Instead of writing entire pipelines from scratch, you can now mix and match building blocks to create custom workflows tailored to your specific needs!

Diffusers 0.36.0: Pipelines galore, new caching method, training scripts, and more 🎄

The release features a number of new image and video pipelines, a new caching method, a new training script, new kernels - powered attention backends, and more. It is quite packed with a lot of new stuff, so make sure you read the release notes fully 🚀 New image pipelines Flux2 : Flux2 is the latest generation of image generation and editing model from...