Skip to content

0.3.1

Latest

Choose a tag to compare

@hattori7243 hattori7243 released this 20 Nov 11:19
43ceac6

Windows and macOS. All three packages are built from 43ceac6 on main. Installation and troubleshooting: see README.md inside each package.

What's Changed

  • Edge0 145B and Edge0 120B stream their experts from the SSD on macOS. On Metal the experts are read on demand into a fixed-size arena sized from the Mac's memory, so both models run on a 64 GB Apple Silicon Mac. (#325)
  • NVFP4 and Q8 on macOS. The NVFP4 folders now run on Metal too, resident or streamed from the SSD, and Edge0 Chat offers all four model folders on macOS. (#246, #319, #325)
  • Q8 D/U by default for Edge0 145B and Edge0 120B on every platform; Q4 stays available. (#325)
  • Edge0 Chat: Settings > Engine > Executable is Built-in or Custom; settings from 0.3.0 or earlier switch to Built-in once, so an upgraded app no longer starts an old engine. Settings > Engine > Context length sets the conversation length (16K to 128K or Custom). The Service page says why a recent request failed or was cut off. (#325)
  • Clearer errors for agent tools. A prompt too long for the context is refused in each API's own error format (OpenAI and Anthropic), which agent and coding tools recognize and handle, for example by compacting their history. (#325)
  • Fixes in edge0-infer and edge0-strata, many contributed by the community. (#218, #237, #238, #247, #290, #291, #300, #308, #312, #315)

Windows x64

edge0-chat-0.3.1-windows-x64.zip

  • Edge0 35B runs on the CPU (AVX2) or on a GPU through Vulkan.
  • Edge0 145B and Edge0 120B run on an NVIDIA GPU (CUDA) of compute capability 7.5 or newer (RTX 20 / 30 / 40 / 50), with the NVIDIA CUDA Toolkit 13 installed.

SHA256: 128cdd9c75e60550ffd1196bef15bf005117d57a65df03e5a434187fead089a8

Windows Arm64

edge0-chat-0.3.1-windows-arm64.zip

  • Edge0 35B runs on the CPU or on a GPU through Vulkan.
  • Edge0 145B and Edge0 120B run on an NVIDIA GPU (CUDA) of compute capability 12.1, such as the NVIDIA RTX Spark N1X, with the NVIDIA CUDA Toolkit 13 for Windows Arm64 installed. The CUDA code is built for sm_121 only.

SHA256: 146442ce9dff822fc345395fb85fad143ff9a1e5c22cdff7bece4cd63d3e3dde

macOS (Apple Silicon)

edge0-chat-0.3.1-macos-arm64.zip

  • macOS 26 or later on Apple Silicon. edge0-infer and edge0-strata are inside Edge0 Chat.app. The app is signed ad hoc and not notarized: on first start, allow it under System Settings > Privacy & Security (see README.md in the package).
  • Edge0 35B runs on Metal or the CPU.
  • Edge0 145B and Edge0 120B run on Metal from their q3k and NVFP4 folders, with the experts streamed from the SSD on a Mac with less memory than the model.

SHA256: 6a48a5e0d6c35b1355125d2eb4dbc0cb07530f972fa906da3dd98280153ac947

Models

Model weights are not included in the packages. Edge0 Chat can download them, or get them here:

Full Changelog: v0.3.0...v0.3.1