DiffusionGemma + Gemma 4 MTP

[danielhanchen]

We've merged over 150 PRs this week so lots of new updates, a new model Hub and look!

DiffusionGemma + Gemma 4 MTP + Audio

• Run and train DiffusionGemma via Unsloth Studio.
• Gemma 4 MTP is here! Run Gemma 4 around 2x faster with MTP - MTP is auto enabled in Unsloth Studio.
• Audio chat is now supported for Gemma 4 (wav, mp3, m4a,flac, webm).

Hub + Download Manager (Experimental)

• Added a new Hub page for browsing, downloading, and managing Hugging Face models and datasets.
• Unsloth can now detect models and datasets already on your machine and show them alongside downloaded assets.
• Downloaded GGUF models now have direct Run / New Chat actions.

Chat with Files / RAG (Experimental)

• Added Chat with Files in Studio, letting you ask questions over your own documents and knowledge bases.
• Supports hybrid search, citations, PDF previews, per-thread documents, and a built-in search_knowledge_base tool.

New Update Button + Hardware Support

• Unsloth now uses fresh, up-to-date llama.cpp prebuilts across CUDA, ROCm, Windows, Linux, and macOS.
• Added an in-app Update llama.cpp button so users can update the local backend without reinstalling Studio.
• Improved Windows / WSL AMD support, Strix Halo ROCm support, Blackwell CUDA selection, and clearer installer messages.

Local Chat, Tools & API Compatibility

• Local tool calling is more reliable, with better ordering of tool cards, fewer duplicate tool loops, and support for tool use with GGUF vision models.
• Improved OpenAI-compatible API and Anthropic-compatible API behavior for local Studio servers, including better errors, token usage, stop reasons, and Claude Code compatibility.

Tool Calling, MCP, Encrypted Cloudflare Tunnels

• Bypass Permissions, Tool Call Permissions (Approve, Always Approve, Deny)
• 50% to 90% less tool call nudging issues without any accuracy loss
• MCP, Artifacts are now select-able
• Tensor parallelism is now enabled for GGUFs - get +30% throughput!
• Cloudflare HTTPS free tunnels is now added allowing for end to end encrypted studios!

Training & General Fixes

• Improved MLX support with better model labels, generation speed stats, and fixes for VLM training.
• Fixed several training and dataset edge cases, including non-writable Hugging Face caches and custom dataset mappings.
• Added many UI polish fixes across chat, menus, model picker, dark mode, import/export, and settings.

To update Unsloth or install a new Unsloth Studio, you must use:

macOS, Linux, WSL:

curl -fsSL https://unsloth.ai/install.sh | sh

Windows:

irm https://unsloth.ai/install.ps1 | iex

Warning

DO NOT USE unsloth studio update since packaging will not get the latest updates

What's Changed

• Studio: llama.cpp update banner redesign, About tab license info, UI polish by @shimmyshimmer in Studio: llama.cpp update banner redesign, About tab license info, UI polish #6196
• Bump install.sh / install.ps1 pin to unsloth>=2026.6.3 by @danielhanchen in Bump install.sh / install.ps1 pin to unsloth>=2026.6.3 #6212
• Expose runtime context length for hub models by @alkinun in Expose runtime context length for hub models #6154
• Studio: fix llama.cpp update banner offering a downgrade / sticking on mix releases by @oobabooga in Studio: fix llama.cpp update banner offering a downgrade / sticking on mix releases #6219
• Fix kwarg spacing in training files to satisfy pre-commit by @shimmyshimmer in Fix kwarg spacing in training files to satisfy pre-commit #6209
• Studio: reword the Cloudflare line when the public probe fails by @danielhanchen in Studio: reword the Cloudflare line when the public probe fails #6217
• fix: deduplicate lemonade ROCm prebuilt selection log by @LeoBorcherding in fix: deduplicate lemonade ROCm prebuilt selection log #6021
• Stop false RoPE 'default' warning and fix rope drift gate on transformers 5 by @danielhanchen in Stop false RoPE 'default' warning and fix rope drift gate on transformers 5 #6223
• fix(studio): load run.py by path for editable installs by @jimdawdy-hub in fix(studio): load run.py by path for editable installs #5909
• fix(studio): inherit llama_extra_args and honor --no-mmproj by @jimdawdy-hub in fix(studio): inherit llama_extra_args and honor --no-mmproj #5902
• fix(studio): adopt server-loaded model before chat auto-load by @jimdawdy-hub in fix(studio): adopt server-loaded model before chat auto-load #5900
• Fix stale sidebar regression test to match the gap-px markup by @danielhanchen in Fix stale sidebar regression test to match the gap-px markup #6232
• Studio: gate the staged prebuilt runtime validation behind a flag (off by default) by @danielhanchen in Studio: gate the staged prebuilt runtime validation behind a flag (off by default) #6216
• Fix FastModel config passthrough for sequence classification by @alkinun in Fix FastModel config passthrough for sequence classification #6203
• fix: decode subprocess output as UTF-8 in save.py on Windows by @dylanschroers in fix: decode subprocess output as UTF-8 in save.py on Windows #6218
• patch: fix EmptyLogits gathering in nested payloads and Accelerate recursively_apply by @MdHussain121 in patch: fix EmptyLogits gathering in nested payloads and Accelerate recursively_apply #6092
• Studio: show Apple GPU temperature and power in the GPU monitor (macOS) by @Ban921 in Studio: show Apple GPU temperature and power in the GPU monitor (macOS) #6187
• Studio: Add inline confirmation (Allow/Always allow/Deny) for tool calls by @oobabooga in Studio: Add inline confirmation (Allow/Always allow/Deny) for tool calls #5869
• Studio: guard Apple GPU power against negative counter-reset readings by @danielhanchen in Studio: guard Apple GPU power against negative counter-reset readings #6235
• Fix step count mismatch when sequence packing is enabled by @IrakliXYZ in Fix step count mismatch when sequence packing is enabled #5967
• fix/uv-bytecode-timeout by @alkinun in fix/uv-bytecode-timeout #6166
• Studio: tune llama.cpp env for data-center GPUs by @danielhanchen in Studio: tune llama.cpp env for data-center GPUs #6098
• Studio: drop the on-disk freshness cache after a llama.cpp update by @danielhanchen in Studio: drop the on-disk freshness cache after a llama.cpp update #6234
• Add missing RAG deps to no-torch Studio runtime requirements by @danielhanchen in Add missing RAG deps to no-torch Studio runtime requirements #6236
• Studio: rounded rectangle hover states for menu items instead of pills by @shimmyshimmer in Studio: rounded rectangle hover states for menu items instead of pills #6210
• docs: repository cleanup by @Agnibha007 in docs: repository cleanup #5617
• Run cross-platform parity test on Windows and macOS in CI by @danielhanchen in Run cross-platform parity test on Windows and macOS in CI #6241
• chore(studio/frontend): normalize line endings to LF by @danielhanchen in chore(studio/frontend): normalize line endings to LF #6012
• fix: respect absolute export paths to prevent cross-drive copy failures (WinError 112) by @anmolxlight in fix: respect absolute export paths to prevent cross-drive copy failures (WinError 112) #6088
• Studio: Add Tensor-Parallel llama.cpp support by @oobabooga in Studio: Add Tensor-Parallel llama.cpp support #6040
• Studio: Add custom provider option to Connections by @Imagineer99 in Studio: Add custom provider option to Connections #6112
• Studio: model selector and settings polish by @shimmyshimmer in Studio: model selector and settings polish #6240
• Studio: login card polish and sidebar label alignment by @shimmyshimmer in Studio: login card polish and sidebar label alignment #6242
• Studio: pinnable plus menu items and saved prompt pins by @shimmyshimmer in Studio: pinnable plus menu items and saved prompt pins #6237
• Studio: bottom update banners, smooth llama.cpp progress, re-prompt after copy by @shimmyshimmer in Studio: bottom update banners, smooth llama.cpp progress, re-prompt after copy #6233
• fix(studio/responses): forward chat_template_kwargs enable_thinking to chat request by @Anai-Guo in fix(studio/responses): forward chat_template_kwargs enable_thinking to chat request #6202
• Studio: fix WSL Strix Halo GPU on reinstall (ROCDXG drop-in + system HIP before bundle) by @danielhanchen in Studio: fix WSL Strix Halo GPU on reinstall (ROCDXG drop-in + system HIP before bundle) #6227
• Studio: fully rounded Hub pills and refreshed menu icons by @shimmyshimmer in Studio: fully rounded Hub pills and refreshed menu icons #6248
• Studio: use px-2.5 for Hub option menu padding by @shimmyshimmer in Studio: use px-2.5 for Hub option menu padding #6249
• Studio: fix Downloaded model list disappearing and order it by last download by @danielhanchen in Studio: fix Downloaded model list disappearing and order it by last download #6247
• Studio: new-chat shortcut, composer draft autosave, archive threads by @NilayYadav in Studio: new-chat shortcut, composer draft autosave, archive threads #5771
• Studio: persist speculative decoding preference across restart and model switch by @oobabooga in Studio: persist speculative decoding preference across restart and model switch #6169
• Studio: refine menu chevron, tick icon, and one-line plus-menu shape by @shimmyshimmer in Studio: refine menu chevron, tick icon, and one-line plus-menu shape #6251
• Studio: serve DiffusionGemma with live in-place denoising and honest stats by @danielhanchen in Studio: serve DiffusionGemma with live in-place denoising and honest stats #6250
• Studio: bundle Gemma 4 chat templates (E2B/E4B + larger) and auto-apply to unsloth/gemma-4-*-GGUF by @danielhanchen in Studio: bundle Gemma 4 chat templates (E2B/E4B + larger) and auto-apply to unsloth/gemma-4-*-GGUF #6245
• feat(studio): implement S3 dataset loading (completes feat(studio): add S3 dataset configuration foundation #5951) by @ashzak in feat(studio): implement S3 dataset loading (completes #5951) #6222
• Studio: cache MCP tool discovery instead of re-probing every chat send by @oobabooga in Studio: cache MCP tool discovery instead of re-probing every chat send #5828
• Fix Studio S3 dataset panel layout by @wasimysaid in Fix Studio S3 dataset panel layout #6252
• Attach DiffusionGemma visual-server from the prebuilt bundle by @danielhanchen in Attach DiffusionGemma visual-server from the prebuilt bundle #6254

New Contributors

• @jimdawdy-hub made their first contribution in fix(studio): load run.py by path for editable installs #5909
• @dylanschroers made their first contribution in fix: decode subprocess output as UTF-8 in save.py on Windows #6218
• @MdHussain121 made their first contribution in patch: fix EmptyLogits gathering in nested payloads and Accelerate recursively_apply #6092
• @Ban921 made their first contribution in Studio: show Apple GPU temperature and power in the GPU monitor (macOS) #6187
• @IrakliXYZ made their first contribution in Fix step count mismatch when sequence packing is enabled #5967
• @Agnibha007 made their first contribution in docs: repository cleanup #5617

Full Changelog: v0.1.451-beta...v0.1.46-beta

---

This discussion was created from the release DiffusionGemma + Gemma 4 MTP.