VeizikNVIDIA and Apple Silicon builds are available now.Release notes →
veiziknative inference engine
Sign inDownload
Product
OverviewWhat the engine isHow it worksWeights in, native execution outPlatformsNVIDIA · Apple SiliconSecurity & licensingDevice-bound, local by designReleasesEvery build, dated
Engineering
All capabilitiesOne researched page per capabilityVideo memory720p that stays under a 24 GB cardA 35B model on one card21 GiB resident on a 24 GB GPUEdge optimizationGenerative AI on a Raspberry Pi CPUIts own stackNo Python, no framework, no cuBLASResume a renderCrash, resume, same result
Models
NVIDIALanguage and video — available nowApple SiliconQwen2.5 · Qwen3 — availableRaspberry Pi · edgeSD1.5 · SDXS · GPT-2 — arriving soonFull catalogueEvery model and its status
Hardware
All hardwareEvery device × every modelPick your deviceWhat runs on yoursNVIDIAAmpere → Blackwell · CUDAApple SiliconM1 – M4 · MetalRaspberry Pi · edgeCPU & NPU — generative AI on a Pi
Benchmarks
All resultsMatrix or pick, by buildBuilds & downloadsEvery build, JSON / CSVMethod & notationHow every figure is taken
Docs
QuickstartInstall to first runCommand referenceEvery documented commandActivate a machineDevices and licenceWhat leaves the machineOffline and networkTroubleshootingWhen something fails
GalleryReleasesSign inDownloadGitHub
Hardware

NVIDIA

CUDA execution on NVIDIA GPUs. Runs on supported NVIDIA generations: Ampere, Ada Lovelace, Blackwell (Hopper untested here).

The NVIDIA packages for Qwen and Wan appear here, with their SHA-256, when their GitHub release page is public.

Overlapping sheets of dark graphite metal, each fold catching a thin warm light seam.

NVIDIA — what runs on it

Supported generations: Ampere, Ada Lovelace, Blackwell (Hopper untested here).

Image rendered by Veizik for NVIDIA, development build (linux x86_64) · SD3.5-medium · NVIDIA RTX 3090 · Powered by Stability AI

Supported combinations

What runs on this target

The download for this target carries these models: Language — Qwen3.5, Qwen3.6; Video — Wan 2.1 · Wan 2.2.

Device memory and a verified context limit are listed per model once a row of its own has been measured on a download; none has been measured on these yet. See the catalogue.