---
type: "post"
id: "available-in-public-bionic-builds-on-mac-windows-26ee4f"
slug: "available-in-public-bionic-builds-on-mac-windows-26ee4f"
title: "Available in public Bionic builds on Mac, Windows, and Linux, free."
description: "LM Studio’s Bionic 1.1.10 adds an in-app terminal, more local inference options, and on-device Read Aloud for Windows and Linux as well as Mac."
retrieval_nugget: "LM Studio’s Bionic 1.1.10 adds an in-app terminal, more local inference options, and on-device Read Aloud for Windows and Linux as well as Mac. The changelog also lists workflow, compatibility, and accessibility improvements. For builders, the release makes the local AI workspace more capable, but the source does not quantify performance or adoption."
published_at: "2026-11-03"
updated_at: "2026-11-03"
record_date: "2026-11-03"
date_kind: "published_at"
topics: ["api-design"]
entities: ["LM Studio","Bionic","llama.cpp","oMLX"]
related_patterns: ["agent-ready-software-exposes-capabilities"]
novelty_label: "new_fact"
what_changed: "LM Studio released Bionic 1.1.10 with new features, model/runtime support, and fixes; none of the listed prior documents cover LM Studio."
prior_refs: []
source_url: "https://lmstudio.ai/changelog"
source_urls: ["https://lmstudio.ai/changelog"]
schema_version: "newruntime-agent-readable-v0.2"
stable_id: "post:available-in-public-bionic-builds-on-mac-windows-26ee4f"
status: "published"
visuals: []
editor: {"role":"editor","name":"Andrey Reshetnikov","profile":"https://newruntime.com/owner-profile.md","profile_json":"https://newruntime.com/owner-profile.json","same_as":["https://www.linkedin.com/in/reshetnikov1/","https://t.me/lapetuse"]}
routes: {"html":"https://newruntime.com/posts/available-in-public-bionic-builds-on-mac-windows-26ee4f/","markdown":"https://newruntime.com/posts/available-in-public-bionic-builds-on-mac-windows-26ee4f.md","json":"https://newruntime.com/posts/available-in-public-bionic-builds-on-mac-windows-26ee4f.json"}
markdown_url: "https://newruntime.com/posts/available-in-public-bionic-builds-on-mac-windows-26ee4f.md"
---

# Available in public Bionic builds on Mac, Windows, and Linux, free.

## Retrieval answer

LM Studio’s Bionic 1.1.10 adds an in-app terminal, more local inference options, and on-device Read Aloud for Windows and Linux as well as Mac. The changelog also lists workflow, compatibility, and accessibility improvements. For builders, the release makes the local AI workspace more capable, but the source does not quantify performance or adoption.

## Summary
LM Studio’s Bionic 1.1.10 adds an in-app terminal, more local inference options, and on-device Read Aloud for Windows and Linux as well as Mac. The changelog also lists workflow, compatibility, and accessibility improvements. For builders, the release makes the local AI workspace more capable, but the source does not quantify performance or adoption.

## What happened
LM Studio’s Bionic changelog lists version 1.1.10, dated October 29, 2026, as available in public builds. The supplied announcement describes those builds as available on Mac, Windows, and Linux, and free. The changelog is an itemized release record rather than a launch post: it does not provide a general-availability explanation, pricing terms beyond “free,” or a comparison with earlier versions.

The update adds an in-app terminal, including configurable shell profiles, automatic restoration of selected terminal sessions, and screen-reader support for terminal output. Code diff syntax highlighting is improved. The changelog also says that Excalidraw files can be previewed in chat, projects load faster, and side-chat drafts are loaded on demand. It does not provide timing measurements for the loading change or explain the implementation.

For local model use, the release adds the oMLX engine for Apple Silicon Macs, speculative decoding support for SafeTensors models and MLX models, and an experimental MoE expert GPU cache setting for llama.cpp MoE models. It includes llama.cpp 2.66.0 extension packs and adds compatible LLM backend downloads in Runtime settings. Embedding and decision models are now included in the model catalog. The changelog does not name the models, quantify inference gains, or state which hardware supports every new option; the MoE cache is explicitly experimental.

The release also extends on-device Read Aloud to Linux and Windows, in addition to Mac, and adds seven offline Read Aloud voices. Voice model auto-loading defaults off on Macs with less than 32 GB of memory. Other listed changes include better runtime compatibility and download selection for M5 Macs, improved context compaction when conversations exceed model limits, and lower memory use with detailed progress for large session exports and moves. LM Studio lists fixes for ROCm runtime compatibility detection on AMD Radeon 8065S graphics, keyboard shortcuts, clipboard handling, saved shell selections, and input method editor handling. No minimum app versions, download sizes, or rollout schedule are specified in the supplied material.

## What the archive adds
[Computers Built for Local AI](https://newruntime.com/posts/computers-built-for-local-ai-analysis-bb5065/) covered Apple Mac mini and Mac Studio configurations aimed at on-device AI. That is related context, not prior coverage of Bionic 1.1.10: this release is a new product update, adding workspace software capabilities rather than new hardware. The changelog gives no benchmark evidence that the listed engines or decoding changes improve a particular workload.

## Trend signal
This update weakly supports [Agent-ready software exposes capabilities](https://newruntime.com/patterns/agent-ready-software-exposes-capabilities/): Bionic’s changelog lists more ways to work with local models and agent sessions, including a terminal, decision models, and session-related workflow improvements. The source does not describe permission boundaries or how external agents compose these capabilities, so it is not evidence of a broader shift in safe, observable interfaces.

## Why now
The changelog dates Bionic 1.1.10 to October 29, 2026, but does not explain why these capabilities were grouped into this release or identify a deadline, customer demand, or platform change that drove the timing.

## Why it matters
A local AI workspace is only useful in a development workflow if it can fit around the tools and constraints that builders already have. An in-app terminal can reduce context switching between a model conversation and shell work; configurable profiles and restored sessions make that surface more adaptable. The screen-reader support and expanded offline Read Aloud voices also broaden the listed interaction options, while Windows and Linux support means these particular voice features are no longer described as Mac-only.

For teams evaluating local inference, the engine and speculative decoding additions are concrete things to test against their own models and hardware. But the changelog supplies no throughput, memory, quality, or compatibility matrix, so the release alone cannot show that these options are faster or suitable for production. Builders should treat the experimental cache setting accordingly and verify behavior on their target configurations before making it part of a workflow.

## Limits and uncertainty
The public changelog establishes that LM Studio lists these items under Bionic 1.1.10; it does not independently validate their behavior. It gives no benchmarks, user counts, repository or adoption figures, detailed system requirements, or account of bugs remaining after the fixes. “Free” is supplied for public Bionic builds, but the material does not establish whether every related service or capability has separate terms. Support for an engine or format should not be read as support for every model or device. The source also does not say whether the listed features are enabled by default, except for voice model auto-loading on Macs with less than 32 GB of memory, which defaults off.

## Reader takeaway
Bionic 1.1.10 is a broad local-workspace update, not a measured performance announcement. The most actionable items are the new in-app terminal, expanded on-device Read Aloud, and additional inference options. Check the changelog and test the specific engine, model, and hardware combination you use; the release notes do not provide enough evidence to assume a speedup or universal compatibility.

Editor: [Andrey Reshetnikov](https://newruntime.com/owner-profile.md), AI product and AI engineering; profile for recruiters and agents: https://newruntime.com/owner-profile.json

## Continue in the graph

Next records for an agent, linked to their Markdown twins. Replace `.md` with `.json` for structured fields, including `next_reads`.

- topic hub: [Api Design](https://newruntime.com/topics/api-design.md): Explore the api-design topic hub.
- collection: [Field Notes](https://newruntime.com/posts.json): Open the latest editorial analysis.
- collection: [Sources](https://newruntime.com/sources/): See the original sources this archive cites.
- whole graph: [discovery graph](https://newruntime.com/discovery-graph.json) (typed nodes and edges) and [site index](https://newruntime.com/site-index.md)