First Release post Unsloth Studio launch!

[shimmyshimmer]

Hey guys, this will be our first release since we launched Unsloth Studio last week. Also from now on you can directly access all our updates through our changelog here: https://unsloth.ai/docs/new/changelog

• You can now update Unsloth! Just use: unsloth studio update
• Tool calling improved. Better llama.cpp parsing, no raw tool markup in chat, faster inference, a new Tool Outputs panel, timers.
• Windows CPU or GPU now works seamlessly. Please reinstall!
• App shortcuts. Once installed, you can now launch in Windows, Macn and Linux via a shortcut icon in the Start / Launch and Desktop.
• Pre-compiled llama.cpp binaries and mamba_ssm for finetuning - 6x faster installs! Also <300MB in size for binaries.
• 50% reduced installation sizes (-7GB or more savings), 2x faster installs and faster resolving. 50% smaller pypi sizes.
• Colab with free T4 GPUs with Unsloth Studio now fixed! Try it here. Due to pre-compiled binaries, it's also 20x faster!
• MacOS and CPU now have Data Recipes enabled with multi-file uploading.
• You can now properly use old GGUFs from Hugging Face or LM Studio
• AMD support preliminary for Linux only machines - auto detects.
• Settings sidebar redesign. Settings are now grouped into Model, Sampling, Tools, and Preferences
• Context length now adjustable. Keep in mind this is not needed as llama.cpp smartly uses the exact context you need via --fit on
• Persistent system prompts and presets. Custom system prompts and chat presets now persist across reloads and page changes.
• Multi-file upload. Data recipes now support multiple drag-and-drop uploads for PDF, DOCX, TXT, and MD, with backend extraction, saved uploads, and improved previews.
• Better chat observability. Studio now shows llama-server timings and usage, a context-window usage bar, and richer source hover cards.
• Better UX overall - clickable links, better LaTeX parsing, tool / code / web tooltips for default cards and much more!
• LiteLLM - Unsloth Studio and Unsloth were NOT affected by the recent LiteLLM compromise. Nemo Data Designer used LiteLLM only up to 1.80, not the affected 1.82.7 or 1.82.8, and has since removed it entirely.
• We now have a new one line install command, just run: Copycurl -fsSL https://unsloth.ai/install.sh | sh

Fixes:

• Windows/setup improvements. Fixed silent Windows exits, Anaconda/conda-forge startup crashes, broken non-NVIDIA Windows installs, and missing early CUDA/stale-venv setup checks.
• System prompts fixed. They work again for non-GGUF text and vision inference.
• GGUF export expanded. Full fine-tunes, not just LoRA/PEFT, can now export to GGUF. Base model resolution is more reliable, and unsupported export options are disabled in the UI.
• Chat scroll/layout fixes. Fixed scroll-position issues during generation, thinking-panel layout shift, and viewport jumps when collapsing reasoning panels.
• Smarter port conflict detection. Studio now detects loopback conflicts, can identify the blocking process when possible, and gives clearer fallback-port messages.

Example of automatic parameter settings for context length etc:

super.final.mp4

What's Changed

• [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci[bot] in [pre-commit.ci] pre-commit autoupdate #4542
• fix: store embedding_learning_rate on self in UnslothTrainingArguments by @GoldenGrapeGentleman in fix: store embedding_learning_rate on self in UnslothTrainingArguments #4531
• studio: persist system prompt and preset settings across navigation by @Imagineer99 in studio: persist system prompt and preset settings across navigation #4538
• studio: stop scroll hijack during generation and fix thinking panel layout shift by @Imagineer99 in studio: stop scroll hijack during generation and fix thinking panel layout shift #4543
• Fix Studio port conflict detection for loopback addresses by @danielhanchen in Fix Studio port conflict detection for loopback addresses #4532
• fix(studio): show Windows-specific reset-password command by @Shine1i in fix(studio): show Windows-specific reset-password command #4529
• fix(studio): restore scroll lock on reasoning panel collapse by @danielhanchen in fix(studio): restore scroll lock on reasoning panel collapse #4545
• fix: always show chat tool icons by @Shine1i in fix: always show chat tool icons #4525
• fix: system prompt ignored in unsloth inference by @Shine1i in fix: system prompt ignored in unsloth inference #4528
• fix: handle prompt/completion datasets in slow-path BOS detection by @danielhanchen in fix: handle prompt/completion datasets in slow-path BOS detection #4548
• fix: give @0xKushwaha git history credit for completion_only_loss fix by @danielhanchen in fix: give @0xKushwaha git history credit for completion_only_loss fix #4552
• ⚠️Remove quarantined litellm for precaution -- Unsloth Studio NOT affected by @danielhanchen in ⚠️Remove quarantined litellm for precaution -- Unsloth Studio **NOT** affected #4553
• fix: pin unsloth>=2026.3.11 in install scripts by @danielhanchen in fix: pin unsloth>=2026.3.11 in install scripts #4556
• Regroup chat settings sidebar into focused sections by @Shine1i in Regroup chat settings sidebar into focused sections #4551
• Add GRPO resume vLLM cleanup guard by @MagellaX in Add GRPO resume vLLM cleanup guard #4411
• fix: prevent UnicodeEncodeError on Windows CP1252 consoles in studio setup by @Krishnachaitanyakc in fix: prevent UnicodeEncodeError on Windows CP1252 consoles in studio setup #4563
• studio: windows desktop shortcut launcher by @Imagineer99 in studio: windows desktop shortcut launcher  #4558
• Remove duplicate frontend assets from wheel (~31 MB savings) by @danielhanchen in Remove duplicate frontend assets from wheel (~31 MB savings) #4567
• feat(studio): training history persistence and past runs viewer by @Shine1i in feat(studio): training history persistence and past runs viewer #4501
• fix: remove auto wandb.finish() after train() to allow post-training evaluate() by @Krishnachaitanyakc in fix: remove auto wandb.finish() after train() to allow post-training evaluate() #4564
• feat: Implement Q-GaLore optimizer and custom embedding learning rate… by @OnePunchMonk in feat: Implement Q-GaLore optimizer and custom embedding learning rate… #4511
• Bump Data Designer to 0.5.4 (removes litellm dependency) by @danielhanchen in Bump Data Designer to 0.5.4 (removes litellm dependency) #4569
• feat(chat): cleaner tool UI, inline LaTeX, clickable links by @Shine1i in feat(chat): cleaner tool UI, inline LaTeX, clickable links #4561
• [Studio] Try installing causal-conv1d from prebuilt wheels if avialable by @Datta0 in [Studio] Try installing causal-conv1d from prebuilt wheels if avialable #4547
• Feature/add dependabot and codeql security checks by @pkloehn1 in Feature/add dependabot and codeql security checks #4479
• build(deps): bump the actions group with 2 updates by @dependabot[bot] in build(deps): bump the actions group with 2 updates #4570
• build(deps): bump oxc-parser from 0.116.0 to 0.121.0 in /studio/backend/core/data_recipe/oxc-validator in the npm-oxc-validator group by @dependabot[bot] in build(deps): bump oxc-parser from 0.116.0 to 0.121.0 in /studio/backend/core/data_recipe/oxc-validator in the npm-oxc-validator group #4571
• Remove advanced CodeQL workflow (conflicts with default setup) by @danielhanchen in Remove advanced CodeQL workflow (conflicts with default setup) #4584
• Add macOS and Linux desktop shortcuts to install.sh by @danielhanchen in Add macOS and Linux desktop shortcuts to install.sh #4568
• perf(studio): upgrade to Vite 8 + auto-install bun for faster frontend builds by @Etherll in perf(studio): upgrade to Vite 8 + auto-install bun for faster frontend builds #4522
• feat(tokenizer): add get_tokenizer_info() diagnostic helper by @cz-03 in feat(tokenizer): add get_tokenizer_info() diagnostic helper #4436
• Add ROCm (AMD GPU) support to studio setup by @danielhanchen in Add ROCm (AMD GPU) support to studio setup #4585
• Consolidate dual venvs and separate install from update by @rolandtannous in Consolidate dual venvs and separate install from update #4530
• studio: stabilize reasoning panel scroll behavior and prevent composer overlap by @Imagineer99 in studio: stabilize reasoning panel scroll behavior and prevent composer overlap #4587
• Use prebuilt llama.cpp for unsloth studio setup by @mmathew23 in Use prebuilt llama.cpp for unsloth studio setup #4562
• fix(studio): add -ngl flag for GPU offloading in llama-server by @danielhanchen in fix(studio): add -ngl flag for GPU offloading in llama-server #4588
• fix(studio): add pip nvidia CUDA libs to LD_LIBRARY_PATH for llama-server by @danielhanchen in fix(studio): add pip nvidia CUDA libs to LD_LIBRARY_PATH for llama-server #4590
• fix(studio): validate bun install and retry from official source on failure by @danielhanchen in fix(studio): validate bun install and retry from official source on failure #4589
• fix(studio): clear bun cache on failure and retry before falling back to npm by @danielhanchen in fix(studio): clear bun cache on failure and retry before falling back to npm #4594
• Pin torch>=2.4,<2.11.0 in Studio installers by @danielhanchen in Pin torch>=2.4,<2.11.0 in Studio installers #4595
• fix(studio): source-build fallback prefers Unsloth's tested tag over upstream latest by @danielhanchen in fix(studio): source-build fallback prefers Unsloth's tested tag over upstream latest #4593
• fix(studio): add bun cache validation to Windows setup.ps1 by @danielhanchen in fix(studio): add bun cache validation to Windows setup.ps1 #4596
• feat: multi-source model discovery (HF default, legacy cache, LM Studio) by @rolandtannous in feat: multi-source model discovery (HF default, legacy cache, LM Studio) #4591
• Add unsloth to User PATH on Windows after install by @danielhanchen in Add unsloth to User PATH on Windows after install #4597
• Add PID file tracking and unsloth studio stop command by @danielhanchen in Add PID file tracking and unsloth studio stop command #4598
• feat(studio): editable context length with Apply/Reset for GGUF settings by @danielhanchen in feat(studio): editable context length with Apply/Reset for GGUF settings #4592

New Contributors

• @MagellaX made their first contribution in Add GRPO resume vLLM cleanup guard #4411
• @Krishnachaitanyakc made their first contribution in fix: prevent UnicodeEncodeError on Windows CP1252 consoles in studio setup #4563
• @OnePunchMonk made their first contribution in feat: Implement Q-GaLore optimizer and custom embedding learning rate… #4511
• @pkloehn1 made their first contribution in Feature/add dependabot and codeql security checks #4479
• @dependabot[bot] made their first contribution in build(deps): bump the actions group with 2 updates #4570
• @cz-03 made their first contribution in feat(tokenizer): add get_tokenizer_info() diagnostic helper #4436

Full Changelog: b8475...v0.1.2-beta

---

This discussion was created from the release First Release post Unsloth Studio launch!.

[Entersjkhdfkjdhfksjf]

Nice work! Any updates on MPS/MLX support?

[shimmyshimmer]

Should be early next month!