Nyra Logic - 7B (Titan Series)
Logihertz Universal Local Intelligence | Production Build v1.0
🛰️ Asset Overview
Nyra Logic is a proprietary neural component engineered by Logihertz Systems (OPC) Private Limited, Mumbai. Designed as a high-fidelity "Local AI" solution for cross-platform deployment, this model leverages High-Fidelity Neural Synthesis to provide deterministic reasoning capabilities optimized for standalone hardware execution without cloud dependency.
📥 Secure Payload Delivery
Format: Safetensors (Master Weights) Access Point: 📁 Access Files & Versions Tab
🏗️ Strategic Ecosystem
☁️ Enterprise Cloud Infrastructure by AWS
Powered by Tier-1 Enterprise Cloud Compute by AWS, enabling massive-scale neural synthesis, high-performance training, and secure model distribution.
🎓 Academic Incubation
Physically incubated at Indala College of Engineering (MOU Partner), sourcing elite Indian engineering talent for proprietary AI research and optimization.
🛡️ Access Control & Security Gating
This repository is a Protected Sovereign Asset. Access is strictly limited to authorized partners and verified institutional requestors.
- Manual Review: All requests are vetted by the Logihertz Directorate (Mumbai R&D Center).
- Turnaround: Vetting is typically completed within 48 Hours.
- Requirement: Requests must provide a valid corporate/academic identity and intended deployment environment.
⚖️ Intellectual Property
Unauthorized weight extraction, redistribution, or reverse-engineering is strictly prohibited under Indian and International Intellectual Property Law. This build is a closed-source production artifact.
Logihertz Systems (OPC) Private Limited
DPIIT Recognized Startup | GST: 27AAGCL6394P1ZL
Mumbai, Maharashtra, India 🇮🇳
© 2026 All Rights Reserved.
- Downloads last month
- 2
Evaluation results
- openai/gsm8k · Gsm8k View evaluation results leaderboard 48
- allenai/ai2_arc · Eval View evaluation results 48
- Rowan/hellaswag · Eval View evaluation results 61.38
- cais/mmlu · Eval View evaluation results 54.33
- truthfulqa/truthful_qa · Eval View evaluation results 47.86
- allenai/winogrande · Eval View evaluation results 62.75
- openai/openai_humaneval · Eval View evaluation results 29.88
- google-research-datasets/mbpp · Eval View evaluation results 5