Model

RogerAI-Apex

rogerai-apex

Our flagship command & language model — engineered for high-stakes enterprise work. Apex reasons across long, complex instructions, digests entire policies and reports, drafts formal communications, and returns structured answers where quality outweighs speed.

Context 32,768Params 80L · 8192hSize 135.43 GB
Download
Model

RogerAI-Core

rogerai-core

The balanced daily workhorse. Core handles Arabic & English conversation, summarization, rewriting, task extraction, and formal drafting — the ideal default model to power internal portals and everyday application workflows.

Context 40,960Params 40L · 5120hSize 27.52 GB
Download
Model

RogerAI-Forge

rogerai-forge

Built for engineers. Forge generates and reviews code across languages, detects bugs, analyzes logs, writes tests, and scripts in PowerShell, Python & SQL — with a massive context window for large repositories and DevOps automation.

Context 262,144Type MoESize 56.89 GB
Download
Model

RogerAI-Index

rogerai-index

The semantic backbone of intelligent search. Index turns text into embeddings, powers meaning-based retrieval, links questions to the right passages, and drives RAG systems across large knowledge bases and archives.

Context 40,960Role EmbeddingsSize 7.50 GB
Download
Model

RogerAI-Rank

rogerai-rank

A lightweight precision layer for retrieval. Rank reorders search results by true relevance, filters out noise, and selects the strongest passages before answer generation — sharpening RAG quality with minimal footprint.

Context 40,960Role RerankerSize 1.12 GB
Download
Model

RogerAI-Scan

rogerai-scan

Vision-language understanding for documents. Scan reads screenshots and scanned pages, extracts data from forms and tables, performs OCR-style comprehension, and describes images and interfaces — perfect for smart archiving and data-entry automation.

Type Vision-LanguageRole OCR / VLSize 16.34 GB
Download

Deployment & Requirements

Recommended baseline for running the RogerAI model family in production.
Operating System
Modern Linux Server
Python
3.10 or 3.11
Core Libraries
torch · transformers · tokenizers · safetensors · accelerate
Storage
NVMe SSD · ≥ 500 GB initial for models, cache & logs
Network
Fast private link between API and model server
Security
Keep model folders private · read-only access to model files
Monitoring
GPU VRAM · RAM · latency · queue length · errors
Serving
Persistent inference service (vLLM · TGI · equivalent)