Login or SignupDownload AppModelsDeveloper DocsLM LinkNewCareersWe're Hiring!
Blog
ChangelogEnterprise SolutionsPrivacy PolicyTerms of Use
← Back to Changelog
May 27, 2025

LM Studio 0.3.16

0.3.16 - Release Notes

New features and improvements in LM Studio 0.3.16:

  • Staff Picks: new aggregate format that combines formats (e.g. GGUF and MLX) into a single entry
  • Custom Settings for models: a new way for model authors to define custom capabilities for models (e.g. Enable Thinking on/off for Qwen3 models)
  • Auto deletion of least recently used Runtime Extension Packs to save disk space
  • Add "Offload KV Cache to GPU Memory" option to model load options and GPU settings
  • Add lms chat command to the CLI to chat with a model in the terminal (Thanks @mayfer!)
  • Ability to use GGUF LLMs as text embedding models (override in My Models > ⚙️)
  • Ability to publish your presets to the LM Studio community hub (see https://lmstudio.ai/docs/app/presets/publish)

Build 8

  • Bugfix: nicer UI for Hugging Face search error when model source doesn't exist
  • Bugfix: avoid showing duplicate models in My Models after clicking "Complete Download"

Build 7

  • Fix prompt processing % UI not showing is User UI mode

Build 6

  • Added a dropdown in the model editor (⚙️ in My Models) to allow overriding the domain type of a model
    • This allows you to use an LLM as a text embedding model, or vice versa
  • Fixed GTE Qwen embedding models are not recognized as text embedding
  • Fixed a bug where if no engine extension pack for a specific format is installed, the indexing will fail
  • Fixed a bug where some strings were parsed as [object Object] when using RAG
  • Fixed a bug showing duplicate models in Load Model in tray menu

Build 5

  • Auto-deletion of least recently used Runtime Extension Packs
    • Can disable in App Settings (⌘/Ctrl + ,)
  • Add tool call logging to "Developer Logs"
  • Fix architecture badge text alignment

Build 4

  • App Settings (⌘/Ctrl + ,): sections now have their own tabs for easier navigation
  • Add button to access downloads panel in "User" UI mode
  • Bugfix: [OpenAI-like REST API] Don't add empty function name string after the first packet of streamed function calls
  • Bugfix: Correctly filter out unsupported models when listing downloaded models
  • Bugfix: Don't assume Vulkan iGPU memory is Shared Memory

Build 3

  • LM Studio REST API (/api/v0): return model capabilities in GET /models response
    • e.g. "capabilities": ["tool_use"]
  • Fix flickering horizontal line at bottom of generating message
  • Fix inability to be create a new folder when no chat is selected
  • Fix prompt processing % UI not showing
  • Fix lms bug that prevents listing downloaded models if an unsupported model is present

Build 2

  • Staff Picks polish: introduce new aggregate format that combines formats (e.g. GGUF and MLX) into a single entry
  • Custom Settings for models
    • A new way for model authors to define custom capabilities for models. For example, Enable Thinking on / off for Qwen3 models
    • At the moment, only available for Qwen3 models under Staff Picks.
      • Later we'll share public docs for doing this with your custom models
  • Add "Offload KV Cache to GPU Memory" option to model load options and GPU settings
  • Add lms chat command to the CLI to chat with a model in the terminal (Thanks @mayfer)
  • Fix nomic-bert-moe, nomic-embed-code not showing under embedding models
  • Show System Prompt button in chat top bar when sidebar is collapsed
  • Use Cmd / Ctrl + Shift + D to create a duplicate of the current chat
  • Use Cmd / Ctrl + E to open the System Prompt editor
  • Use Cmd / Ctrl + W to close the System Prompt editor tab when it is open
  • Fix SmolVLM jinja render error "Expected iterable type in for loop: got StringValue"

Build 1

  • Public Preview: ability to publish your presets to the LM Studio community hub! 👾
    • Docs: https://lmstudio.ai/docs/app/presets/publish
    • Downloading public presets does NOT require a login. Publishing, liking, forking does
    • Share links with others directly, or on Discord / Reddit / X (No way to browse presets yet)
    • This is a new feature, we'd love to get your feedback: bugs@lmstudio.ai
  • Add "Reveal in Finder" context menu option on the chat sidebar body
    • Open your conversations folder in your OS native file explorer and reorg files / folders
      • Changes will be reflected live in LM Studio
  • Fix Qwen3 jinja template parsing error "OpenSquareBracket !== CloseStatement"
  • [MLX] Register chat_template.jinja as a source for chat templates
  • [MLX] Fix installation issue that caused ModuleNotFoundError: No module named 'mlx_engine'
    • You can press the "Fix" button in the runtimes page if you are seeing this error with an existing MLX runtime