https://aitoolchain.io/tools/lm-studio New releases and features in LM Studio, tracked by The AI Toolchain. en Fri, 28 Aug 2026 14:11:35 GMT https://lmstudio.ai/changelog/lmstudio https://lmstudio.ai/changelog/lmstudio Fri, 28 Aug 2026 14:11:35 GMT LM Studio 0.4.22 adds DFlash, DSpark, and MTP assistant drafters plus tool-returned image support in OpenAI and Anthropic APIs. • Adds support for images returned by tools in OpenAI-compatible and Anthropic-compatible chat APIs. https://lmstudio.ai/changelog/lmstudio/lmstudio-v0.4.22 https://lmstudio.ai/changelog/lmstudio/lmstudio-v0.4.22 Fri, 28 Aug 2026 00:00:00 GMT LM Studio 0.4.22 adds DFlash/DSpark/MTP model support and tool-returned image support across OpenAI and Anthropic API endpoints. • Supports images returned by tools in OpenAI-compatible `/v1/responses` and `/v1/chat/completions` endpoints, and Anthropic-compatible `/v1/messages` endpoint — enabling multimodal tool-call workflows. • Adds support for DFlash, DSpark, and MTP assistant drafter model architectures (requires llama.cpp engine version 2.29.1 or newer). • Enables in-app updates for Linux AppImage and `.deb` installations. https://lmstudio.ai/changelog/lmstudio/lmstudio-v0.4.21 https://lmstudio.ai/changelog/lmstudio/lmstudio-v0.4.21 Wed, 12 Aug 2026 00:00:00 GMT LM Studio 0.4.21 adds local server API key support for enterprise network mode and updated llama.cpp advanced load settings. • Supports local server API key authentication when using enterprise internal network endpoint mode. • Updates advanced load settings for `mmap`, `mlock`, and `direct io` for llama.cpp engines 2.28.1 and newer. https://lmstudio.ai/changelog/lmstudio/lmstudio-v0.4.20 https://lmstudio.ai/changelog/lmstudio/lmstudio-v0.4.20 Wed, 22 Jul 2026 00:00:00 GMT LM Studio 0.4.20 adds enterprise network model endpoints, Bionic LM Link support, and a Llama.cpp log-level developer setting. • Adds a developer setting to control the Llama.cpp engine log level. • Adds support for enterprise internal network model endpoints. • Adds support for using local device models in Bionic over LM Link via the LM Studio Engine Protocol.