ThumbLLM 0.1.0 — Ornith 1.5 9B CPU Edition

ThumbLLM — Ornith 1.5 9B CPU Edition

Run Ornith 1.5 9B locally on a Windows x64 CPU with no discrete GPU required.

This ThumbLLM edition packages a tested model, quantization, llama.cpp runtime, and CPU inference configuration into a portable local AI application.

Configuration

• Model: OBLITERATUS/Ornith-1.5-9B-OBLITERATED
• Quantization: Q4_K_M
• Format: GGUF
• Backend: CPU
• Platform: Windows
• Architecture: x64
• Runtime: llama.cpp
• llama.cpp build: b10603
• llama.cpp commit: c060ca974

Features

• Built-in desktop chat
• Local OpenAI-compatible API
• Automatic model download
• SHA-256 model verification
• Streaming generation
• Live inference statistics
• CPU-only inference
• Offline inference after initial model download

Installation

1. Download the ThumbLLM .exe.
2. Place it in a folder of your choice.
3. Double-click the executable.
4. ThumbLLM will download the required GGUF model if needed.
5. The local chat interface and API will start automatically.

No llama.cpp installation or model configuration is required.

Recipe

See the full tested configuration in:

recipes/Ornith-1.5-9B-OBLITERATED-Q4_K_M-CPU-Win-x64/README.md