unslothai/unsloth v0.1.902-beta on GitHub

This release brings faster navigation, shareable run settings, and clearer errors to Unsloth Desktop. It also 4x speeds up Laya decisions, expands hosted Decision API support, and keeps NVFP4, INT4, and MXFP4 checkpoints in 4-bit during LoRA training.

Highlights

Command palette on Cmd/Ctrl+P. Also, you can now share your GGUF model's run settings
Laya decisions up to 4.1x faster, plus connections to TypeSafe and other hosted providers
Clearer errors that explain failed loads and generations, with a View logs button
NVFP4 LoRA with 45% lower peak memory on Qwen3.8-27B-NVFP4
INT4 + MXFP4 checkpoints stay packed during LoRA, including Kimi-K2.7-Code and gpt-oss

Desktop + chat

Cmd/Ctrl+P opens a command palette, and GGUF run settings can be shared as a link or saved without loading.
Continue response extends a finished reply, and Fix with the model puts HTML canvas errors in the message box.
A warning flags models that only partly fit on the GPU, and the Images gallery loads thumbnails first.
The desktop app adds Cmd/Ctrl + and - zoom, Windows 11 style buttons on Windows and Linux, and a check before Update stops training.

Lower-memory checkpoint training

Pre-quantized 4-bit checkpoints now train on their exact published weights. Training reads them packed, so a loaded model needs about as much memory as its 4-bit checkpoint.

Qwen3.8-27B-NVFP4 peaks at 40.2 GB instead of 72.9 GB on one RTX PRO 6000, with 11% shorter steps.
Dense compressed-tensors INT4 checkpoints train as published on NVIDIA GPUs. Qwen3-8B w4a16 peaks at 8.5 GB, not 18.8 GB, on a B200.
Kimi-K2.7-Code trains on 6 B200s with its INT4 experts packed, where a 16-bit copy would need about 2 TB.
gpt-oss loaded with load_in_4bit = False keeps its MXFP4 experts, so gpt-oss-120b trains on one B200.

How it works

Triton kernels dequantize each layer on the fly, bit for bit matching compressed-tensors, and backward keeps only the packed weights.
INT4 stays exact instead of being re-quantized to NF4, which rounds every weight twice and raised Qwen3-8B perplexity by 1.6%.
For inference-only gpt-oss, UNSLOTH_MXFP4_KEEP_PACKED=0 restores the native load, which prefills faster.

Decision API + Laya

Laya decision models now answer short requests up to 4.1x faster, and hosted decision models work too.

Add TypeSafe or another hosted decision provider under Connections, then pick its models in Settings > API.
With the Decision API on, enable Unsloth Decisions in the chat's MCP menu for a decide tool.
On a B200, a 3-question guardrail check takes 2.8 ms instead of 11.6 ms. Long inputs run at about the same speed.
Laya loads in about 9 s instead of 17 s and needs about half the peak host RAM.

Image + video

Image models keep INT8 or FP8 when offloading. On a 10 GiB-capped B200, Z-Image takes 2.3 s, not 38.4 s.
Qwen-Image-2.1 renders 1536x1536 images up to 3.6x faster on a Radeon 8060S.

Training + models

Fine-tune T5, T5Gemma, BART, Marian and other text encoder-decoders with FastModel.
The new unsloth eval command runs lm-evaluation-harness tasks on a checkpoint or LoRA adapter.
Gemma 4 31B training and batched generation now work on models split across GPUs.

AMD + Apple Silicon

On Windows, RX 5000 through RX 9000 cards get PyTorch from AMD's multi-arch index.
MLX moves to mlx 0.32.3 and mlx-vlm 0.7.4, and quantized-KV vision loads can batch replies.
On a Mac, export LoRA adapters as PEFT or GGUF for GPU machines.

Download Unsloth Desktop

Unsloth Desktop is free and open source. Download it for:

Platform	Link
Windows	Download
macOS	Download
Linux x64 / Ubuntu (deb)	Download
Linux ARM64 / Ubuntu 24.04+ (deb)	Download
Linux x64 (AppImage)	Download
Windows ARM64	Download

What's Changed

Bump install.sh / install.ps1 pin to unsloth>=2026.9.12, unsloth-zoo>=2026.9.8 by @danielhanchen in #12209
Give the Kimi processor test's stand-in tokenizer convert_tokens_to_ids by @danielhanchen in #12206
Follow chat export calls in the cancelled-save contract instead of a fixed file list by @danielhanchen in #12203
Add Idefics3 support (Granite Docling VLM) by @danielhanchen in #4241
Repair the checks red on main after today's Mistral-format, GKD, laya, llm-compressor and HF token merges by @danielhanchen in #12218
Load compressed-tensors packed INT4 checkpoints for QLoRA without a 16-bit copy (exact packed INT4 by default, NF4 fallback), and train remote DeepSeek-V3 MoE code (Kimi-K2.7-Code) by @danielhanchen in #11537
fix(studio): prevent terminal tool hangs in credential scan (#12048) by @Souravrajvi0 in #12087
Default loader.py's Mistral-format names in the precision-conflict harness by @danielhanchen in #12229
Studio: say why a generation or a model load failed, and offer the log by @danielhanchen in #8804
Train Kimi-K3 with its MXFP4 experts kept packed and dequantized on the fly by @danielhanchen in #11750
fix: align GKD Liger loss with the native objective by @taking-lying-flat in #12007
Studio: resolve repo-relative image paths in the simple image+text VLM converter by @breken-ai in #11897
fix(studio): reuse hardware monitor drag and resize handling for API monitor by @Imagineer99 in #12204
Unsloth Studio (AMD/Windows): real torchao INT8 / FP8 export on Windows ROCm by @danielhanchen in #7102
Studio: clamp unsupported reasoning effort to offered levels by @oobabooga in #12208
Studio: load Laya without randomly initialising its vocabulary embedding by @danielhanchen in #12210
Scope the runtime encoding scan past vendored laya while its loader supplies UTF-8 by @danielhanchen in #12230
Studio: load Images strip tiles as thumbnails instead of full PNGs by @oobabooga in #12228
Images: smaller prompt box radius, inset scrollbar, taller default by @shimmyshimmer in #12223
Library delete dialog: no dark mode border, cancel on outside click by @shimmyshimmer in #12222
Studio: laya download confirm by @NilayYadav in #12196
feat(studio): default the model source to ModelScope where Hugging Face is restricted by @Lyxot in #12197
Studio: fix prompt timestamp layout, accessibility, and imported times by @shimmyshimmer in #12173
Studio: explain failed and recurring llama.cpp runtime repairs by @umran666 in #12106
Studio: faster Laya inference (marker-only head, CUDA graphs) by @danielhanchen in #12224
Library: equal size chat cards and no loading placeholders by @shimmyshimmer in #12221
Run the user message timestamp Playwright check in Frontend CI by @danielhanchen in #12255
Studio: say when a model only partly fits on the GPU by @danielhanchen in #8812
Studio: push only the exported merged model to the Hub on a Mac by @NilayYadav in #12239
Fix sidebar More menu hover and collapsed border hover by @Sneakr in #12216
Studio: keep earlier messages when you send audio by @NilayYadav in #12244
Studio: show the model a phone photo the right way up by @NilayYadav in #12241
Studio: name the llama.cpp graph-scheduler abort and stop reloading into it by @danielhanchen in #6741
Fix NaN gradients from compiled flex attention on short static training batches by @danielhanchen in #12207
Studio: stop chats with images from blocking other chats on GGUF models by @NilayYadav in #12236
Studio: stop tool calls from getting stuck in Running forever by @NilayYadav in #12234
Studio: read a remote image URL on safetensors and MLX models instead of ignoring it by @NilayYadav in #12243
Composer attachments: first page previews, Library icons, lighter cards by @shimmyshimmer in #12261
Keep import unsloth working when vLLM needs transformers 5 by @danielhanchen in #12268
Fall back to PEFT's LoRA forward for float32 base weights by @danielhanchen in #7020
Reuse the dY @ B.t() product in the fused LoRA backward by @danielhanchen in #12212
Studio: let the chat model use the Decision API through MCP by @NilayYadav in #12232
Read the llama.cpp changelog after it settles, by content rather than innerText by @danielhanchen in #12273
Average GRPO gradients across DDP ranks, and keep MiniCPM3's pre-head scaling by @danielhanchen in #12193
Add Qwen2.5 Coder model support to registry and chat templates by @danielhanchen in #4221
Studio: do not fail a cached dataset load on a bookkeeping write by @BardiaKoopah in #8057
Wait for the model-config Load to finish before the test moves on by @danielhanchen in #12270
Scale the attachment preview text with the UI font size by @danielhanchen in #12276
Unsloth Studio: explain why Model Memory skips the RAM lock when a model is fully on the GPU by @danielhanchen in #9571
Studio: title a chat with the connection that answered it, not the local model by @Lyxot in #9106
Fix packed and grouped Q2 GGUF variant detection by @Imagineer99 in #7033
Improve empty text dataset validation by @Imagineer99 in #6646
Studio: fit the context again when MTP is forced on a reload by @NilayYadav in #12231
Studio: show an error when Gemini fails a reply partway through by @NilayYadav in #12242
ci: enforce npm ci on Studio install paths (follow-up to #5604) by @danielhanchen in #5616
Studio: let the dataset check accept chat columns that training already supports by @NilayYadav in #11849
Unsloth Studio / Desktop: allow text-encoder nvfp4 on pre-Blackwell GPUs and explain TE precision refusals by @mahiatlinux in #11413
refactor(trainer): add warning for ignored eval_steps by @danielhanchen in #4225
fix: handle zero-strided tensors in fast_rope_embedding (#3781) by @danielhanchen in #4233
Studio: hide the stray scrollbar under the composer at fractional zoom by @shimmyshimmer in #12279
Unsloth Studio / Desktop (AMD): "Automatic" llama.cpp follows the installed torch on a mixed NVIDIA+AMD host by @LeoBorcherding in #12250
Unsloth Studio / Desktop (AMD): explain why an NVIDIA+AMD host installed with CUDA_VISIBLE_DEVICES="" shows no GPU by @LeoBorcherding in #12249
Studio desktop: let the sidebar shrink to 224px by @shimmyshimmer in #12283
Studio: reopen the artifact panel after it has been dragged shut by @danielhanchen in #12278
Studio: say when an attached PDF has no readable text by @NilayYadav in #12237
Studio: read attached HTML files in the encoding they declare by @NilayYadav in #12235
Decline the fused LoRA kernels under FSDP, where the weights are shard views by @danielhanchen in #11132
Studio: allow long inputs for embedding models by @NilayYadav in #11721
feat(mlx): support DDP in CLI training by @Lyxot in #7166
Studio: save the chat template a base model was trained with by @NilayYadav in #12240
Unsloth Studio (AMD RDNA 1, Windows): install ROCm torch for the RX 5700 XT from AMD's multi-arch index by @LeoBorcherding in #11755
Studio: keep pop-up messages off the Run settings panel by @NilayYadav in #11486
Fix DDP "marked ready twice" for VLMs with CPU offload + TiledMLP by @danielhanchen in #4240
Studio: show download progress during desktop updates by @NilayYadav in #11291
Unsloth Studio / Desktop: add Canvas's console errors, and offer them to the model by @LeoBorcherding in #11744
fix(prompt storage): follow up, restore prompt list bookmarking, the kind pill and drag autoscroll by @LeoBorcherding in #9788
Raise the minimum typer version so the unsloth command starts by @NilayYadav in #11148
Unsloth Studio: run fp16 Laya checkpoints in fp16 on MLX by @Lyxot in #12256
Qwen3.5: skip compiled regions on eager decode steps, make CUDA graph decode opt-in by @danielhanchen in #12213
Studio: stop the stall watchdog from killing Xet downloads that are still receiving data by @NilayYadav in #11594
Studio: keep what you typed with a document when a long chat is compacted by @NilayYadav in #12238
Train LoRA on NVFP4 compressed-tensors checkpoints with the weights kept packed (W4A16) by @danielhanchen in #12185
Studio: keep working when GitHub rate limits or is down by @NilayYadav in #10670
Studio tests: forget the managed provider URL setting cache around each test by @danielhanchen in #12288
Unsloth Studio / Desktop: show the GPU llama.cpp runs on when it is CUDA or ROCm and differs from training by @LeoBorcherding in #12251
Unsloth Studio / Desktop: support custom vision projectors by @mahiatlinux in #11352
Studio: Respect the thinking budget on /v1/messages by @NilayYadav in #11724
Keep a model's no-placement parameters on CPU (Qwen3.8-Flash-Next n-gram table) by @danielhanchen in #12141
Studio: keep unsloth start alive while a LoRA adapter's base model downloads by @NilayYadav in #10550
Baseline the thirteen unsloth-zoo 2026.9.8 findings after review by @danielhanchen in #12292
Studio: show attachments when editing a message by @NilayYadav in #11054
Unsloth Studio installer (AMD/Windows): install PyTorch for RX 5000 to 9000 cards from AMD's multi-arch index by @LeoBorcherding in #11846
Fix: Support past_key_values in model.generate for multi-turn conversations by @danielhanchen in #4232
Studio: command palette (Cmd/Ctrl+P) by @NilayYadav in #6846
Unsloth Studio / Desktop: reload llama.cpp models on reconnect by @mahiatlinux in #11341
Harden timing-dependent Studio backend tests (nvidia-smi cache, cold /api/health) by @danielhanchen in #12275
Let FastModel and FastLanguageModel load the same family in one process, in either order by @danielhanchen in #12219
Studio: save a model's run settings without loading it by @NilayYadav in #10216
Keep LoRA on dense Linears sharing a name with fused MoE experts in PEFT's v5 conversion by @danielhanchen in #12281
Train gpt-oss MXFP4 LoRA through unsloth-zoo's packed experts when load_in_16bit is not set by @danielhanchen in #11929
feat(studio): reuse MLX VLM prompt snapshots in the resident vision batch by @Lyxot in #12263
Let Gemma 4 31B train when it is split across GPUs by @NilayYadav in #12233
Preserve repo id case when resolving a model name by @BardiaKoopah in #8058
feat(studio): Mac adapter-format option for LoRA export, with GGUF adapters on macOS by @Lyxot in #7539
Studio: list large MCP tool schemas compactly and load the full schema on demand by @NilayYadav in #11046
Support for Seq2Seq Models (T5, T5Gemma, etc.) by @danielhanchen in #4226
Studio: run large Laya requests in token-budgeted chunks by @danielhanchen in #12271
Studio: stabilize generalized compare mode (clean replacement) by @Imagineer99 in #7407
Add unsloth eval CLI command by @NilayYadav in #6824
Reload LoRA adapters trained with added tokens by @danielhanchen in #4219
Unsloth Studio / Desktop: expose prefill progress in API monitor by @mahiatlinux in #11161
Studio: one top row and layout fixes for the Windows desktop app by @mahiatlinux in #12070
Studio desktop: Cmd/Ctrl + and - zoom with a zoom popup by @shimmyshimmer in #12280
Studio: use the Hugeicons internet icon for every globe by @shimmyshimmer in #12321
Studio: use FileEmpty02Icon for generic file chips by @shimmyshimmer in #12324
Studio: mark adjusted backup timestamps as estimated by @shimmyshimmer in #12322
Studio: stop polling focus while message menus are open by @shimmyshimmer in #12323
Studio desktop tests: keep temp paths distinct when the clock repeats by @danielhanchen in #12296
Studio tests: check the compare composer's paste path, not one spelling of it by @danielhanchen in #12325
Tests: let the TiledMLP DDP workers import unsloth_zoo on a CPU runner by @danielhanchen in #12329
Studio: use Refresh01Icon and FileEmpty02Icon everywhere by @shimmyshimmer in #12331
Studio: Library grid cards keep one icon position and use the full width by @shimmyshimmer in #12330
Studio: tidy the Skills dialog rows and fields by @shimmyshimmer in #12332
Studio: Doc01 icon for Word and Google Docs files by @shimmyshimmer in #12340
Studio: lift Library grid icons to the card's middle by @shimmyshimmer in #12341
Studio: more edge padding on the model picker dropdowns by @shimmyshimmer in #12338
Studio: open the sidebar More menu on hover again by @shimmyshimmer in #12339
Studio: use the standard chevrons in place of Hugeicons' curved ones by @shimmyshimmer in #12337
Studio: stop normal chat from opening the API monitor by @Imagineer99 in #12305
Studio: fix Java initialization and home in the Linux tool sandbox by @oobabooga in #12294
Studio: don't re-prompt a finished code answer on safetensors and MLX models by @NilayYadav in #12309
Studio: keep answer columns out of the system prompt by @NilayYadav in #12306
unsloth start pi and dsh: stop cutting every reply at 8,192 tokens by @NilayYadav in #12315
Studio: flag responses that appear to stop while quoting a token by @oobabooga in #12259
Studio: keep web search source links for unsloth start claude by @NilayYadav in #12307
Move the attention mask to each layer's device in the fast decode loops by @danielhanchen in #12290
fix(studio): mark CSV exports as UTF-8 so Excel shows non-Latin text by @Mathews-Tom in #11971
Studio: keep code indentation when an HTML file is attached to a chat by @NilayYadav in #12312
Audio: keep PyAV decoding working on PyAV 19, and stop the tests needing an AMR encoder by @danielhanchen in #12295
Studio: accept multiple audio files per message by @shimmyshimmer in #12267
Parallel-isolation guard: exempt the nvidia-smi fake's poll deadline by @danielhanchen in #12353
Studio: ask before Update stops a training run in the desktop app by @NilayYadav in #12313
Validate unsloth train flags the way the config file is validated by @Abhishek-B-R in #11780
Fix three main CI failures from the Mac adapter export and seq2seq PRs by @danielhanchen in #12354
Studio frontend test: find the More flyout by its props in any order by @danielhanchen in #12356
Studio: chunk large Qwen-Image-2.1 attention queries on ROCm by @oobabooga in #12335
Studio desktop: survive AppKit exceptions during event dispatch on macOS by @shimmyshimmer in #12277
Studio: fix the faint seam beside the chat composer by @Imagineer99 in #12304
feat(studio): batch KV-quantized and TurboQuant MLX loads with prompt snapshot reuse by @Lyxot in #12343
Studio: load plain FP8 encoder safetensors without torchao by @oobabooga in #12333
Studio: let the Mac window be moved during an app update by @NilayYadav in #12316
Studio: keep menus and popovers below the desktop titlebar by @shimmyshimmer in #12348
fix(studio): record the bit widths of prequantized MLX models by @Mathews-Tom in #11940
Fix the ShareGPT mapping for Llama 3.1, Qwen and Gemma chat templates by @NilayYadav in #12314
Studio: report CUDA graphs as on once the deferred speed profile arms them by @danielhanchen in #12301
Keep packed INT4 layers off the fused inference kernel in train mode by @danielhanchen in #12349
Studio: isolate the Windows Terminal on MXC's default tier by @Etherll in #12328
Studio: prompt caching for OpenRouter by @Lyxot in #12264
Force non-reentrant gradient checkpointing for DeepSeek-V4.1 by @danielhanchen in #12318
Studio: count tokens for templates that refuse an empty chat by @shimmyshimmer in #12336
Studio: catch the tensor split abort behind a gdb backtrace by @oobabooga in #12293
Keep frozen BatchNorm running stats fixed during LoRA training by @danielhanchen in #12319
Widen the pinned child GPU mask when extra args name a companion device (#11810) by @deepspace28 in #11823
Studio: turn train on completions back on when leaving CPT by @NilayYadav in #12308
Studio: fix Qwen-Image-2.1 recompiles and Ideogram 4 attention and device errors on torch 2.11 to 2.14 by @danielhanchen in #12350
Studio: fix the desktop island corner and border seam by @Imagineer99 in #12285
Move the Unsloth Studio MLX pins to mlx 0.32.3 and mlx-vlm 0.7.4 by @Lyxot in #12334
Honour FP8Linear.block_size in the patched FP8 forward (32x32 block checkpoints) by @danielhanchen in #12317
Write the Ollama Modelfile from the trained chat template by @NilayYadav in #12311
Fix MiniMax-H3 compiled render on torch 2.12, 2.13 and 2.14 by @danielhanchen in #12326
Studio: load LTX-2 and LTX-2.3 on the pinned transformers 5.5 by @danielhanchen in #12299
scan_packages: refuse VCS, URL and local-path specs before pip download by @danielhanchen in #12357
Studio: let GGUF vision models see dark text in transparent images by @NilayYadav in #12310
Keep masked head_dim 256 SDPA training off cuDNN attention on SM100 (torch 2.14 NaN grads) by @danielhanchen in #12344
Unsloth Studio (AMD): run native image generation on an NVIDIA card next to ROCm torch by @LeoBorcherding in #12252
fix(lora): preserve all active adapters with PEFT fallbacks by @taking-lying-flat in #12214
Studio: load downloaded tokenizers offline, size unsloth mirrors from the family table by @danielhanchen in #12300
Build the marlin_gemm call from the op schema so packed INT4 inference works on vLLM 0.29 by @danielhanchen in #12320
Studio video: keep an explicit fast LTX-2.3 single-file load resident when it fits by @danielhanchen in #12345
Studio: keep MiniMax-H3's int8 denoiser and compile when it has to offload by @danielhanchen in #12286
Model-config UI test: stop waiting on a request whose finish event never arrives by @danielhanchen in #12363
Studio: keep INT8 and compile on conventional video models when they have to offload by @danielhanchen in #12289
Studio frontend test: drain the previous runtime's writes before each reasoning-effort scenario by @danielhanchen in #12367
Baseline the moved transformers 5.18.0 testing_utils polling-loop site after review by @danielhanchen in #12368
Studio: keep int8 / fp8 image transformers quantised when the load has to offload by @danielhanchen in #12287
Studio: style canvas notices like toasts by @shimmyshimmer in #12358
Studio: keep the model name ahead of its format and quant when the header is tight by @shimmyshimmer in #12396
Studio: compact context usage ring when the chat header is squeezed by @shimmyshimmer in #12355
fix(device_type): flush and fence through the shared device helpers by @li-lizhe in #12284
Studio: open a skill when its row chevron is clicked by @NilayYadav in #12375
Studio: keep the answers typed into a PDF form when it is attached to a chat by @NilayYadav in #12377
Studio: stop offering Claude sampling settings that are silently ignored by @NilayYadav in #12383
Studio: count companion assets in the Hub download size of image and video GGUFs by @oobabooga in #12370
Feat share model run settings through links by @Sneakr in #11710
studio: make media family overrides structural by @Imagineer99 in #10150
Studio: keep and Vec text in chat replies by @NilayYadav in #12376
Harden installer source selection, ROCm helper staging and npm scanner cleanup by @danielhanchen in #12402
Remove xFormers built for another torch after Linux repair by @tweekli in #11783
truststore: keep TLS verification on when handshakes overlap on one context by @danielhanchen in #12403
Read a sign-flipped SVD basis as the same subspace in the Q-GaLore schedule by @danielhanchen in #12399
Create the SentencePiece scratch directory without a check-then-create race by @danielhanchen in #12400
Only turn on HF_HUB_ENABLE_HF_TRANSFER in synthetic.py when hf_transfer is installed by @danielhanchen in #12406
Studio: train an uploaded CSV's NA, None and 00501 cells as written by @NilayYadav in #12379
Studio: fix Base vs LoRA compare for voice messages on fine-tuned Whisper by @NilayYadav in #12381
Compare Kaggle reference repo ids without case and prefetch the Qwen3 4bit repo under its new spelling by @danielhanchen in #12411
Seed preview: check path-backed image cells against the account's workspace by @danielhanchen in #12404
Studio: say Pin in model menus, and drop the border on right-click submenus by @shimmyshimmer in #12414
Stop the Chat UI persisted-monitor reset from running script in the stale page by @danielhanchen in #12412
Studio: keep a local model's built-in system prompt when the date setting is on by @NilayYadav in #12382
Studio: let the Decision API use TypeSafe, Liquid AI, OpenRouter and other System One servers by @NilayYadav in #12373
Keep the MLX adamw_8bit optimizer instead of collapsing it to adamw by @vineethsaivs in #9950
Treat {{ and }} in a merged prompt as literal braces by @vineethsaivs in #10742
fix(studio): keep scoped download progress tied to current files by @Imagineer99 in #12390
Studio: continue finished replies and resume GGUF reasoning by @oobabooga in #12371
Studio: keep Qwen-Image-2.1 image conditioning finite on ROCm by @oobabooga in #12360
Restore full-rank gradients after Q-GaLore updates by @vineethsaivs in #10878
Studio: fix PDF previews after a PDF attachment is extracted by @shimmyshimmer in #12346
Studio: stop the backend crashing on startup when memory is tight by @NilayYadav in #12374
Studio: warn that a Web share link from a remote Studio exposes its address by @danielhanchen in #12413
Studio: smaller scroll to bottom button, visible in dark mode, with a setting to hide it by @shimmyshimmer in #12417
Stop conversation_extension crashing when the caller keeps their columns by @vineethsaivs in #8373
Fix desktop icon clarity with supplied artwork and Tauri resource icons by @wasimysaid in #12352
Studio: always show composer attachments as cards, rename sent layouts by @shimmyshimmer in #12418
Studio: remove the Projects section setting from Chat settings by @shimmyshimmer in #12421
Studio: keep the arrow cursor on a sent prompt's time by @shimmyshimmer in #12420
Keep an explicit HF_HUB_ENABLE_HF_TRANSFER and install hf_transfer in Core CI by @danielhanchen in #12394
Studio: remove the white seam above the chat panel on Windows dark mode by @danielhanchen in #12425
Studio: avoid slow cold MIOpen searches on gfx1151 by @oobabooga in #12359
Studio: faster MiniMax-H3 GGUF renders (resident under memory auto, sd.cpp pin upgrade, speed_mode=max kernels) by @danielhanchen in #12405
Studio: keep Word footnotes and endnotes in chats and knowledge bases by @NilayYadav in #12378
Studio: make Compare in Chat load the full fine-tune that just finished by @NilayYadav in #12380
Stop the model selector's format and quant suffix clipping descenders by @danielhanchen in #12427
fix: allow configuring Pi output token limit by @Imagineer99 in #12393
Studio: keep the sidebar's bottom fade in step with the list by @shimmyshimmer in #12423
Studio: keep the 1024 canvas when auto precision picked the Qwen-Image-2.1 quant by @danielhanchen in #12428
Studio: move Blender MCP setup out of Manage MCP servers into the composer by @NilayYadav in #12430
Studio: explain the all-columns-dropped recipe error in UI terms by @danielhanchen in #12416
Studio: Qwen-Image-2.1 placement from measured sizes, int8 under offload on torchao 0.17, balanced fit check by @danielhanchen in #12408
Read the desktop New chat button contract token by token by @danielhanchen in #12426
feat(studio): configurable RAG upload extensions via RAG_UPLOAD_EXTS by @Souravrajvi0 in #11499
Unsloth Studio: return the spoken text from /audio/generate instead of a truncated label by @LeoBorcherding in #12386
Studio: keep PyTorch mirror leaves out of query tokens by @alkinun in #10546
perf(studio): reuse embeddings for identical files in linked folders by @Mathews-Tom in #11972
Studio: render MCP Apps widgets in the chat thread by @NilayYadav in #9301
Bump install.sh / install.ps1 pins to unsloth>=2026.9.13, unsloth-zoo>=2026.9.9 by @danielhanchen in #12432
fix(studio): use resolved public id in embeddings/completions monitor by @Souravrajvi0 in #9346
Studio: keep rounded boxes rounded when they scroll by @shimmyshimmer in #12431
Studio: browse temporary Linux mounts under /media and /mnt by @Souravrajvi0 in #10048
Make the registry's default quant_type usable by @vineethsaivs in #9887
Size the packed attention mask at the padded length, not the token count by @vineethsaivs in #8278
Keep the GRPO eval batch a multiple of num_generations by @vineethsaivs in #12362
Prefetch Tauri NSIS and WebView2 tools before the Windows desktop build by @danielhanchen in #12434
fix(studio): retry custom gateways with max_completion_tokens after max_tokens 400 (#10787) by @Souravrajvi0 in #10837
Collapse the llama.cpp install branch that never branched by @vineethsaivs in #10268
Resolve a sentence-transformers modules.json module class through the same trust gate as upstream by @danielhanchen in #12436

New Contributors

@Mathews-Tom made their first contribution in #11971

Full Changelog: v0.1.900-beta...v0.1.902-beta

unslothai/unsloth v0.1.902-beta Command Palette + Desktop UI/UX on GitHub