# transformers v5.15.0 — Release: v5.15.0 - Product: transformers (https://whatsnew.fyi/product/transformers) - Vendor: huggingface - Date: 2026-08-10 - Version: v5.15.0 - Original notes: https://github.com/huggingface/transformers/releases/tag/v5.15.0 - Permalink: https://whatsnew.fyi/product/transformers/releases/v5.15.0 What's New is an index, not a publisher: every entry below links to the vendor's own release notes, which are the authoritative source. Entries are labelled where they are hand-curated sample data, pre-releases, or drawn from a secondary source such as a developer blog. Reuse: the summaries, labels and curation here are © What's New. Quote freely with attribution and a link back; wholesale republication of the corpus is not permitted — terms: https://whatsnew.fyi/terms. The vendors' own release notes remain their publishers'. --- - **added** — Add Muse Glimmer multimodal model support - **added** — Add Granite-swa and Granitemoe-swa model support - **added** — Add A.X-K1 model support - **added** — Add A.X-K2 model support - **added** — Add Cosmos3 Edge model support - **changed** — Kernels are now opt-in rather than mandatory for linear attention models - **changed** — Cache cropping API now only accepts negative values instead of absolute sizes - **changed** — Enable SDPA and other attention backends for T5 and propagate to the T5 family - **removed** — Remove private helper functions from multimodal processor files - **fixed** — Fix multi-head latent attention (MLA) cache compression - **fixed** — Optimize Flash Attention max sequence length computation in vision models - **fixed** — Fix BlockMask crash in CTRL flex-attention generation - **fixed** — Fix sdpa prefill with position bias - **fixed** — Fix dtype alignment in Kosmos2/Kosmos2_5 embedding merges - **fixed** — Fix position-embedding initialization fallback in Phi4Multimodal - **fixed** — Resolve PIL resize parity in Hunyuan-VL - **fixed** — Patch stop-sequence handling in the image-text-to-text pipeline - **changed** — Refactor linear attention models for better maintainability - **changed** — Make Gemma 4 heterogeneous attention config explicit via per_layer_config - **changed** — Improve MPS support via metal-flash-sdpa integration #### 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 assistants, Claw- or Hermes-like setups. Muse Glimmer is a dense 30B parameter model consisting of: - 2B ViT-style encoder for vision (Perception Encoder) - 28B parameter text decoder We're covering it in the following blogpost: http://hf.co/blog/muse-glimmer image --- ###### GraniteMoeSWA & GraniteSWA image **Links:** [Documentation](https://huggingface.co/docs/transformers/main/en/model_doc/granitemoe_swa) * Add Granite-swa and Granitemoe-swa model support (#47179) by @daviswer in [#47179](https://github.com/huggingface/transformers/pull/47179) **Links:** [Documentation](https://huggingface.co/docs/transformers/main/en/model_doc/granite_swa) * Add Granite-swa and Granitemoe-swa model support (#47179) by @daviswer in [#47179](https://github.com/huggingface/transformers/pull/47179) --- ###### A.X-K1 & A.X-K2 image **Links:** [Documentation](https://huggingface.co/docs/transformers/main/en/model_doc/axk2) * Add AXK2 from SKT (#47528) by @vasqu in [#47528](https://github.com/huggingface/transformers/pull/47528) **Links:** [Documentation](https://huggingface.co/docs/transformers/main/en/model_doc/axk1) * add_axk1 (#46867) by @kmswin1 in [#46867](https://github.com/huggingface/transformers/pull/46867) --- ###### Cosmos3 Edge image **Links:** [Documentation](https://huggingface.co/docs/transformers/main/en/model_doc/cosmos3_edge) * Add Cosmos3 Edge model support (#47181) by @atharvajoshi10 in [#47181](https://github.com/huggingface/transformers/pull/47181) ##### Breaking changes Kernels are now opt-in rather than mandatory for linear attention models (Mamba, GDN, Conv-only, etc.), so users who relied on automatic kernel selection must explicitly enable kernels to maintain previous behavior. * 🚨 [`Kernels`] Refactor all linear attn models & native kernels fallback (#47630) by @vasqu The cache cropping API now only accepts negative values (relative offsets) instead of absolute sizes, so users calling crop methods directly must update their code to pass negative values accordingly. * 🚨 [cache] Cropping can only be done with negative values (#47720) by @Cyrilvallez T5 and its model family (MT5, LongT5, etc.) now support SDPA and other attention backends via `ALL_ATTENTION_FUNCTIONS`, meaning the default attention implementation may change and users relying on the previous eager-only path should explicitly set `attn_implementation="eager"` if needed. * 🚨 Enable SDPA (and other attention backends) for T5 and propagate to the T5 family (#47014) by @jiqing-feng Several small private helper functions (e.g., `_is_url`, `_build_image_tokens`) have been removed from multimodal processor files, so users or downstream libraries that imported these private functions directly must remove or replace those references. * :rotating_light: Processors update the rest (#46556) by @zucchini-nlp ##### Attention This release includes several attention fixes and improvements, including correcting Multi-Head Latent Attention (MLA) cache compression, optimizing Flash Attention max sequence length computation _[Truncated at 4000 characters — full notes: https://github.com/huggingface/transformers/releases/tag/v5.15.0]_