v0.1.39-beta

**_v0.1.39-beta bug fix_**
**May 5th 2026** Fixes chat history not being shown (existing chat history is not lost) and attachments not attaching correctly. The bug was render-only - use `2026.5.2` or directly call `curl -fsSL https://unsloth.ai/install.sh | sh` or `unsloth studio update` to update

Run local LLMs with tools like [Claude Code](https://unsloth.ai/docs/basics/claude-code) and [Codex](https://unsloth.ai/docs/basics/codex) by connecting them to Unsloth’s API endpoint. This lets you run models like [Qwen](https://unsloth.ai/docs/models/qwen3.6) and [Gemma](https://unsloth.ai/docs/models/gemma-4) locally, with additional features such as self-healing tool calling, code execution, and web search. Unsloth makes it easy to deploy a fast API inference endpoint that provides:
- [Self-healing tool calling](https://unsloth.ai/docs/new/studio/chat#auto-healing-tool-calling), which helps reduce broken or malformed tool calls by 50%
- [Code execution](https://unsloth.ai/docs/new/studio/chat#code-execution) support, allowing Bash and Python execution for more accurate code outputs.
- Advanced [Web search](https://unsloth.ai/docs/new/studio/chat#advanced-web-search) that visits and actually reads webpages to gather in-depth info.
- [Automatic inference settings](https://unsloth.ai/docs/new/studio/chat#auto-parameter-tuning) for GGUF models (temp, top-k etc.)
<br><img width="350" alt="change profile settings" src="https://3215535692-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FxhOjnexMCB3dmuQFQ2Zq%2Fuploads%2F1s98Id9xclzwMfxjXw2O%2Funsloth%20api%20cropped.png?alt=media&token=64fac263-ca5b-4447-a740-41f58ec94904" />

Models loaded in Unsloth (including GGUFs) are exposed as an authenticated API via `llama-server`. A long API key is generated for security reasons like how OpenAI provides one. Your local models can then be used directly in your preferred AI agent, SDK, or chat client. Unsloth speaks two dialects on the same port:

- Anthropic-compatible `/v1/messages`  for Claude Code, OpenClaw, the Anthropic SDK, and any client that expects the Messages API.
- OpenAI-compatible `/v1/chat/completions` and `/v1/responses` for the OpenAI SDK, OpenCode, Cursor, Continue, Cline, Open WebUI, SillyTavern, and any OpenAI-compatible tool.
- Both support streaming, tool calling (OpenAI tools / Anthropic tools), and vision inputs.

## New models
We've also got a handful of new models to run including NVIDIA [Nemotron 3 Nano Omni](https://unsloth.ai/docs/models/nemotron-3-nano-omni), IBM [Granite 4.1](https://unsloth.ai/docs/models/ibm-granite-4.1) and [Mistral 3.5](https://unsloth.ai/docs/models/mistral-3.5) Medium. We helped Mistral solve some issues with implementation in transformers and GGUFs.

## Unsloth Updates
- Stopped Studio training runs can now resume from checkpoints.
- Chat threads now autosave and persist more reliably.
- DPO training hangs in multi-process setups were fixed.
- VLM GRPO support improved with MROPE updates.
- Studio’s stop button now properly stops generation.
- Fix chat template disappearing after browser refresh

## What's Changed in Unsloth
* Studio: use (gguf) context length before max seq length by @G07cha in https://github.com/unslothai/unsloth/pull/5111
* chore: fix typo cleanup across tests and backend strings by @luojiyin1987 in https://github.com/unslothai/unsloth/pull/5152
* fix: guard resolve_model_class fallback against unresolvable transformers AutoModel entries by @Etherll in https://github.com/unslothai/unsloth/pull/5155
* Studio: kill in-flight llama-server before spawning a new one by @danielhanchen in https://github.com/unslothai/unsloth/pull/5171
* Studio: stop currency escape from breaking inline LaTeX by @danielhanchen in https://github.com/unslothai/unsloth/pull/5170
* Studio: probe AMD GPUs in llama-server VRAM detection by @danielhanchen in https://github.com/unslothai/unsloth/pull/5172
* Studio: make stop button actually stop generation by @danielhanchen in https://github.com/unslothai/unsloth/pull/5069
* Studio: add github_repo seed reader and GitHub Support Bot recipe by @danielhanchen in https://github.com/unslothai/unsloth/pull/5169
* fix(studio): use endswith for mmproj F16 variant selection by @LeoBorcherding in https://github.com/unslothai/unsloth/pull/5184
* [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci[bot] in https://github.com/unslothai/unsloth/pull/5204
* Fix Windows install when paths contain spaces or Python 3.14 is on PATH by @Etherll in https://github.com/unslothai/unsloth/pull/5201
* Studio: Preserve transparency in uploaded profile avatars by @Imagineer99 in https://github.com/unslothai/unsloth/pull/5200
* UX: single chat header error placement and selector alignment by @Imagineer99 in https://github.com/unslothai/unsloth/pull/5173
* Studio: Refine chat preset and group built-in presets by @Imagineer99 in https://github.com/unslothai/unsloth/pull/5159
* Studio: Fix image-only chat requests failing validation by @Imagineer99 in https://github.com/unslothai/unsloth/pull/5212
* Studio: fix 7 failing studio_unit_tests on main by @danielhanchen in https://github.com/unslothai/unsloth/pull/5216
* Patch checkpoint reload init functions to strip unsupported args by @Datta0 in https://github.com/unslothai/unsloth/pull/5167
* Studio: Fix clipped model selector text descenders by @Imagineer99 in https://github.com/unslothai/unsloth/pull/5210
* Fix DPO trainer multi process hang by @Datta0 in https://github.com/unslothai/unsloth/pull/5199
* Studio: Pin assistant-ui core for fresh installs by @Imagineer99 in https://github.com/unslothai/unsloth/pull/5229
* Fix local model scanner to handle ollama cloud models by @Anish9901 in https://github.com/unslothai/unsloth/pull/5220
* Fix Studio desktop tray installer and titlebar and bux fixes by @wasimysaid in https://github.com/unslothai/unsloth/pull/5179
* MROPE for VLM GRPO by @Datta0 in https://github.com/unslothai/unsloth/pull/5198
* install: overlay unsloth-zoo from git main on --local by @rolandtannous in https://github.com/unslothai/unsloth/pull/5242
* Studio: Fix chat template disappearing after browser refresh by @Imagineer99 in https://github.com/unslothai/unsloth/pull/5209
* studio: add --local to setup.sh + overlay unsloth-zoo from git main by @rolandtannous in https://github.com/unslothai/unsloth/pull/5252
* Fix/windowsprebuilt by @mmathew23 in https://github.com/unslothai/unsloth/pull/5241
* Studio: Add dataset upload dropzone and update preserve think copy by @Imagineer99 in https://github.com/unslothai/unsloth/pull/5253
* Add Qwen3.6 support by @rolandtannous in https://github.com/unslothai/unsloth/pull/5257
* Studio: Chat thread autosave persistence by @Imagineer99 in https://github.com/unslothai/unsloth/pull/5256
* Studio: Enable deleting fine-tuned chat models by @Imagineer99 in https://github.com/unslothai/unsloth/pull/5234
* Studio: Add checkpoint resume for stopped training runs by @Imagineer99 in https://github.com/unslothai/unsloth/pull/5255
* Studio: Polish spacing and profile input radius by @Imagineer99 in https://github.com/unslothai/unsloth/pull/5222
* Fix check for libcurl headers in install.sh by @LFd3v in https://github.com/unslothai/unsloth/pull/5251
* Default Studio host to 127.0.0.1 and prompt before auto-start by @rolandtannous in https://github.com/unslothai/unsloth/pull/5267
* Studio: forward llama-server args from `unsloth studio run` , activate `unsloth run` , and allow passing model:quant to load models by @rolandtannous in https://github.com/unslothai/unsloth/pull/5271
* Studio: Always show API usage examples and docs links by @Imagineer99 in https://github.com/unslothai/unsloth/pull/5270
* Studio: Change API Keys settings to API Access by @Imagineer99 in https://github.com/unslothai/unsloth/pull/5268
* unsloth run: add --enable-tools/--disable-tools server-side tool policy by @rolandtannous in https://github.com/unslothai/unsloth/pull/5277
* fix: use % 8 instead of // 8 in FP8 weight shape check by @Ricardo-M-L in https://github.com/unslothai/unsloth/pull/5243
* Pin Studio GGUF export to llama.cpp's local convert script by @mmathew23 in https://github.com/unslothai/unsloth/pull/5275
* fix KVCache estimates for gemma4 style sliding window models by @Datta0 in https://github.com/unslothai/unsloth/pull/5225
* Update VRAM estimator to cater to broader model configs by @Datta0 in https://github.com/unslothai/unsloth/pull/5175
* Fix FastSentenceTransformer loading with newer sentence-transformers by @Etherll in https://github.com/unslothai/unsloth/pull/5259
* Studio: Preserve chat history during autosave by @Imagineer99 in https://github.com/unslothai/unsloth/pull/5278

## What's changed in Unsloth-Zoo
* Fix fused CE grad scaling under DDP by @danielhanchen in https://github.com/unslothai/unsloth-zoo/pull/434
* Fused CE backward: guard scaling=0, drop tensor path, use out-of-place mul by @mmathew23 in https://github.com/unslothai/unsloth-zoo/pull/610
* Fix/gemma4moefix by @mmathew23 in https://github.com/unslothai/unsloth-zoo/pull/612
* MROPE for VLM GRPO by @Datta0 in https://github.com/unslothai/unsloth-zoo/pull/614
* Double-buffer GPU activations for overlapping H2D copy with backward compute by @ruixiang63 in https://github.com/unslothai/unsloth-zoo/pull/534
* fix(temporary_patches/utils): add missing comma in __all__ (raise_error / Unpack) by @Anai-Guo in https://github.com/unslothai/unsloth-zoo/pull/617
* Fix qwen lora extractor for diff peft versions by @Datta0 in https://github.com/unslothai/unsloth-zoo/pull/618
* fix: use backend device type in GGUF merge path by @andomeder in https://github.com/unslothai/unsloth-zoo/pull/615
* Add unsloth_compiled_cache to gitignore by @Datta0 in https://github.com/unslothai/unsloth-zoo/pull/622
* Allow local convert_hf_to_gguf.py via UNSLOTH_LLAMA_CPP_SCRIPTS_DIR by @mmathew23 in https://github.com/unslothai/unsloth-zoo/pull/621

**Full Changelog**: https://github.com/unslothai/unsloth/compare/v0.1.37-beta...v0.1.38-beta