Artificial intelligence (AI) is becoming central to telecom networks, but general-purpose AI models often struggle with industry-specific tasks. Seeking to address that gap, AT&T has unveiled OTel 2.0, an open AI model built specifically for telecom operators.
The company said the model is designed to improve the accuracy of network-related tasks while significantly lowering AI inference costs, making large-scale AI deployments more practical for telecom providers.
Also Read: NVIDIA Unveils 24/7 AI Agents for Telecom Networks as Industry Pushes Towards Autonomous Operations
Built For Telecom From the Ground Up
Unlike conventional AI models trained on broad internet data, OTel 2.0 has been developed using telecom-focused information.
The model is a post-trained version of Google’s Gemma 4 31B-IT and was trained with more than 400 billion telecom-specific tokens, selected from over one trillion processed tokens.
AT&T developed the model in collaboration with the GSMA, while AMD, Microsoft Azure, Dell Technologies and Red Hat supported the training effort. According to the GSMA, OTel 2.0 currently ranks at the top of the Open Telco AI benchmark leaderboard and is the largest open model developed specifically for the telecom industry.
Also Read: Airtel Priority Postpaid 5G Push: Fresh Telecom Battle Triggered Over Net Neutrality
Focus on Lower Costs and Higher Productivity
AT&T said reducing the cost of running AI is as important as improving model performance. The company revealed that its AI systems now process an average of 45 billion tokens every day.
To manage those workloads efficiently, it has introduced an AI Gateway that automatically routes requests to the most suitable AI model based on cost and performance. The company claims this approach has reduced AI inference costs by up to 90%, resulting in savings worth millions of dollars while maintaining service quality.
OTel 2.0 is intended to complement this strategy by offering a specialised model that can run on appropriately sized hardware, whether deployed on-premises or in the cloud.
Also Read: TRAI Proposes Penalties on Telecom Firms Over Poor Complaint Resolution
Industry-wide open AI initiative
The launch builds on the GSMA’s Open Telco AI initiative introduced earlier this year to encourage operators, vendors and researchers to develop telecom-grade AI models using open standards and shared datasets.
AT&T said OTel 2.0 has been released as an open model so the wider telecom industry can further fine-tune it for applications such as network troubleshooting, product development, network configuration and customer service automation.
The company believes specialised AI models will become increasingly important as telecom operators seek to automate network operations while keeping computing costs under control.
Samarjit Kaur is a journalist and communications professional covering technology & emerging digital trends. With a focus on clarity and context, she reports on developments shaping industries and governance. When not reporting, she chooses to plug-in and relax on her playlists and plan her next bucket-list trips!
