LM Studio Developers
@lmstudiodevs
Updates for developers building with @lmstudio SDKs and APIs 👾 npm i @lmstudio/sdk
LM Studio 0.3.31 has shipped! What's new: 🏞️ OCR, VLM performance improvements 🛠️ MiniMax-M2 tool calling support ⚡️ Flash Attention on by default for CUDA 🚂 New CLI command: `lms runtime` See it in action 👇
LM Studio queues requests for a single loaded model instance, not allowing true parallelism (at least for now). THE TRICK : If you have sufficient VRAM, you can load the same model multiple times under different names (instances). Simultaneous requests to these instances will…
LM Studio now ships for NVIDIA's DGX Spark! @nvidia DGX Spark is a tiny but mighty Linux ARM box with 128GB of unified memory. Grace Blackwell architecture. CUDA 13. ✨👾
In addition to the venerable chat completions compat API, @lmstudio now supports /v1/responses! 1. Swap out the openai base url to point to LM Studio 2. Load up gpt-oss 3. Profit
Introducing OpenAI Responses API compatibility! /v1/responses on localhost. Supports stateful responses, custom tool use, and setting reasoning level for local LLMs. 👇🧵
LM Studio 0.3.28 is out now 🛳️ 🫰 Easily choose MLX and GGUF variants, or different quantizations of the same model!
You can just run qwen3-coder on a macbook w/ @lmstudio
qwen3-coder is so shockingly solid in cline when I run it locally on my 36gb ram macbook
Preparing for a packed 2 weeks of updates
LM Studio 0.3.27 build 1 (beta) is available now. New: when loading a model, the selected context length will be taken into account for memory guardrails estimates
LM Studio CLI tool to manage, automate, and script local LLM workflows from the terminal
There's a new open embedding model in town! lms get google/embedding-gemma-300m 300m parameters, 2048 context length, supports 100+ languages.
Introducing EmbeddingGemma: our new open, state-of-the-art embedding model designed for on-device AI 📱
.@cline now has a new “compact system prompt”, designed for local models. Thoughtful context construction is key when using local models as coding agents. Try it with Qwen3 Coder 30B: lms get qwen/qwen3-coder-30b
United States Trends
- 1. Pond 223K posts
- 2. Kim Davis 5,743 posts
- 3. #IDontWantToOverreactBUT N/A
- 4. Semper Fi 8,195 posts
- 5. Marines 40K posts
- 6. Go Birds 7,104 posts
- 7. #MYNZ N/A
- 8. #MondayMotivation 42.4K posts
- 9. Obergefell 3,859 posts
- 10. Veterans Day 21.5K posts
- 11. Edmund Fitzgerald 6,657 posts
- 12. #5SOS_SELFIEDAY N/A
- 13. Obamacare 213K posts
- 14. Good Monday 50.8K posts
- 15. Victory Monday 3,609 posts
- 16. #USMC 1,820 posts
- 17. Ken Burns N/A
- 18. Talus Labs 26.4K posts
- 19. Correísmo Nunca Más N/A
- 20. Rudy Giuliani 33.3K posts
Something went wrong.
Something went wrong.