Personal AI Launches MODEL-4 for Network-Delivered Personal Intelligence

SAN FRANCISCO, Oct. 20, 2026 /CourierPR/ -- Personal AI has unveiled MODEL-4, an operating system designed to deliver personal intelligence directly into the network, enabling telecommunications operators to provide intelligence as a network service to every subscriber. This innovative platform, which integrates with NVIDIA, Microsoft, Rafay, Cisco, and Telegent, aims to transform how AI is delivered and utilized on a global scale.

MODEL-4 is positioned to bring intelligence to the network edge, leveraging the unique advantages that telecommunications operators possess. These operators can connect nearly every individual and business worldwide, ensuring a trusted and secure identity for every line. With MODEL-4, operators can now offer personal intelligence alongside traditional services like talk, text, and data, billed in a manner that aligns with existing plans.

The new system combines proprietary memory, unique identity, on-network inference, and multi-modal agents into a service delivered on every line, marking a significant shift in how AI is integrated into everyday communication.

The core capabilities of MODEL-4 include personal memory, network identity, on-network small language models (SLMs), voice and text agents, and token plans. Personal memory provides a persistent model for each subscriber, allowing the system to remember, learn, and evolve over time. Network identity ties the subscriber's SIM and phone number to the agent's identity, eliminating the need for separate apps or logins. On-network SLMs handle about 95% of traffic inside the operator's network, routing only complex requests to larger models for greater efficiency and better economics.

Voice and text agents enable subscribers to utilize agents through the communication channels they already use, including native calls and texts. Token plans allow operators to package and bill intelligence alongside existing talk, text, and data plans, generating new revenue streams. The system is designed to deliver 16 ms time-to-first-token and 405 milliseconds end-to-end voice latency, with a cost of $0.02 per million tokens, and operator gross margins exceeding 90% at scale.

NVIDIA provides the accelerated computing infrastructure that enables MODEL-4 to move AI inference from centralized data centers into the network, closer to where subscribers live and work. The platform utilizes Multi-Instance GPU (MIG) partitioning on NVIDIA RTX PRO 6000 Blackwell servers (96 GB VRAM) to isolate workloads, hosting up to 40 distilled small language model instances alongside ASR and TTS lanes per GPU slice. This ensures predictable sub-second latency and zero cross-tenant interference.

Microsoft Azure Local extends Azure capabilities to infrastructure the operator runs in its own network, allowing operators to keep subscriber memory and data on-premises, under the same controls that govern their other subscriber information. Rafay provides infrastructure orchestration that automates how MODEL-4 is provisioned and operated across operator and edge environments, reducing deployment time from years to weeks. Cisco optimizes the network for intelligence as a new class of traffic, while Telegent unifies U.S. carrier voice, messaging, and cellular with the communications businesses already run on.

In summary, MODEL-4 represents a significant step forward in the delivery of personal intelligence as a network service, leveraging the unique position of telecommunications operators to provide a new category of value to their subscribers.