37 tok/s from Qwen3-Next-80B (48 GB) on a 32 GB laptop, 24 without drafts vs 2.8 for llama.cpp; reading 62 vs 16 tok/s. Faster than llama.cpp on all 13 models compared in Q4_K_M. LLMs bigger than your RAM, pure C, CPU or GPU.
C avx2ccpu-inferenceggufinference-enginelinuxllmlocal-llmmixture-of-expertsmoenvmequantizationqwen3simdspeculative-decodingvulkan
Release 2026-10-13 (5): vkcheck in the binary packages for Linux and Windows
camauri4df49705622f
diff --git a/.github/workflows/release.yml b/.github/workflows/release.ymlindex 3f3cc9a..b8def8b 100644--- a/.github/workflows/release.yml+++ b/.github/workflows/release.yml@@ -70,7 +70,7 @@ jobs:P="$PACKAGE/janas"mkdir -p "$P/services" "$P/lib" "$P/include/janas" "$P/docs"for f in janas-chat janas-server janas-get janas-bench janas-mcp \- gguf2jns jns_planes hf2jns_mtp jns_check; do+ gguf2jns jns_planes hf2jns_mtp jns_check vkcheck; doinstall -m 755 "$B/$f" "$P/"donefor s in src/services/*/; do@@ -213,7 +213,7 @@ jobs:P="$WPACKAGE/janas"mkdir -p "$P/services" "$P/lib" "$P/include/janas" "$P/docs"for f in janas-chat janas-server janas-get janas-bench janas-mcp \- gguf2jns jns_planes hf2jns_mtp jns_check; do+ gguf2jns jns_planes hf2jns_mtp jns_check vkcheck; doinstall -m 755 "$B/$f.exe" "$P/"donefor s in src/services/*/; dodiff --git a/dist/README-windows.md b/dist/README-windows.mdindex 4962d12..75cef78 100644--- a/dist/README-windows.md+++ b/dist/README-windows.md@@ -53,6 +53,7 @@ The first says which models it knows, their size and the memory they want; the s| `jns_planes` | cuts the experts of a converted mixture into bit planes, so that fewer bits can be read when the memory is short || `hf2jns_mtp` | makes a model's multi-token prediction block from its original checkpoint || `jns_check` | what a `.jns` file says about itself: architecture, sizes, types, licence |+| `vkcheck` | whether this computer's GPU can run Janas's GPU backend, and why not: what the backend needs with the driver's answer, then three short tests on the GPU || `services/` | Janas's services, which bring data of the world to a model: flights, weather, maps, prices, earthquakes, Wikipedia, GitHub, git, this computer. `janas-chat` starts them by itself; any other MCP client can use them. See `services/README.md` || `lib/`, `include/` | `libjanas_llm.dll` and `libjanas_mcp.dll`, their import libraries for MinGW (`.dll.a`), their C headers and FreeBASIC bindings, for programs of your own; the programs above do not need them |diff --git a/dist/README.md b/dist/README.mdindex 0e3084d..af7a86d 100644--- a/dist/README.md+++ b/dist/README.md@@ -47,6 +47,7 @@ Janas-LLM, the inference engine of Janas, runs large language models on ordinary| `jns_planes` | cuts the experts of a converted mixture into bit planes, so that fewer bits can be read when the memory is short || `hf2jns_mtp` | makes a model's multi-token prediction block from its original checkpoint || `jns_check` | what a `.jns` file says about itself: architecture, sizes, types, licence |+| `vkcheck` | whether this computer's GPU can run Janas's GPU backend, and why not: what the backend needs with the driver's answer, then three short tests on the GPU || `services/` | Janas's services, which bring data of the world to a model: flights, weather, maps, prices, earthquakes, Wikipedia, GitHub, git, this computer. `janas-chat` starts them by itself; any other MCP client can use them. See `services/README.md` || `lib/`, `include/` | `libjanas_llm` and `libjanas_mcp` with their C headers and FreeBASIC bindings, for programs of your own; the programs above do not need them |