GPT-OSS 120B Ultra Reasoning

GPT-OSS 120B Ultra is the most powerful reasoning module in the Eidolon ecosystem. With 120 billion parameters, it delivers performance comparable to cloud-based frontier models — running entirely on your hardware, with full privacy.

Built for professionals who need maximum reasoning capability without sending sensitive data to the cloud: legal analysis, medical literature review, code architecture, research synthesis.

Features (both configurations)

• 🧠 120B parameter reasoning engine
• 📄 Document reading and analysis (PDF, Word, text)
• 📋 Markdown chat export and printing
• 👤 PIF (Personal Information Framework) — context-aware responses
• 🔗 Moltbook social module integration
• 💬 Full Eidolon Chat interface
• 🧠 Long-term memory — persistent context across sessions
• 🌐 Internet search with citations
• ☁️ Cloud API bridge — optional connection to external models (OpenAI, Anthropic, Google and more) via API key

Configuration: LITE Requirements: 64–95 GB RAM, CPU-only (no dedicated GPU required)

• All features above
• ❌ Voice interaction not available
• ❌ Vision (image analysis) not available

Configuration: FULL Requirements: 96 GB+ RAM, GPU with 16+ GB VRAM

• All features above
• ✅ Voice interaction (speech-to-text + text-to-speech)
• ✅ Vision — image and document visual analysis via Gemma 4B sidecar

Use Cases

• Legal document analysis and synthesis
• Medical literature review
• Complex code architecture planning
• Academic research synthesis
• Long-form reasoning tasks

System requirements:

• LITE: Windows 10/11, 64 GB RAM minimum, 60 GB free disk space
• FULL: Windows 10/11, 96 GB RAM, GPU with 16+ GB VRAM, 60 GB free disk space