Model library

Open models compiled by plow for specific hardware. Run one with plowrt on your own machine, or deploy it across a fleet with InferOS.

plowrt 0.1.3 One line, Linux and macOS

curl -fsSL https://pdist.infervisor.ai/install.sh | sh

Install

Gemma 4 31B gemma-4-31b

The largest dense Gemma 4, instruction tuned, in bf16 and FP8 on NVIDIA.

Gemma 4 E4B gemma-4-e4b

Google's small instruction-tuned Gemma 4, the text model of the voice-agent stack.

LLM 4B effective (8B with per-layer embeddings) 128k native context bf16

Orpheus 3B orpheus

Canopy Labs' Orpheus 3B, a Llama-based codec LM with a SNAC decoder, served with the same codec-LM contract as Veena.

Text to speech 3B 128k native context bf16

Qwen3.8 27B qwen3.8-27b

Qwen3.8-27B with gated DeltaNet recurrent layers, in bf16 and FP8 on NVIDIA.

Qwen3 0.6B qwen3-0.6b

Qwen's smallest Qwen3, published for Apple Silicon.

Qwen3 ASR 0.6B qwen3-asr-0.6b

The 0.6B Qwen3 ASR model, for latency-first transcription on a small GPU.

Speech to text 0.6B 64k native context bf16

Veena veena

Maya Research's Veena, a Llama-style codec language model with a SNAC decoder for Hindi and English speech.

Text to speech 3B 128k native context bf16