MeetKai and GSMA Unveil ModelChorus at MWC26 Barcelona for AI Language Evaluation

MeetKai and GSMA Launch ModelChorus at MWC26 Barcelona

At the Mobile World Congress 2026 (MWC26) in Barcelona, MeetKai Inc. has partnered with the GSMA to unveil an innovative platform known as ModelChorus. This online tool is designed to empower mobile operators, model developers, and ecosystem partners in evaluating AI language models, with a focus on enhancing the accessibility and utilization of language-ready AI in under-resourced languages and regional dialects.

This launch could not have come at a more pivotal moment in the tech industry. With language diversity becoming increasingly essential in AI applications, ModelChorus stands out by providing a mechanism for users to perform real-time evaluations and comparisons of different AI models. Visitors attending MWC26 can engage with ModelChorus firsthand at the GSMA Pavilion in Hall 4, Stand 4F30, or explore its functionalities via the alpha version at www.modelchorus.com.

Why ModelChorus?

ModelChorus aims to set a new standard for how AI models are evaluated in practical settings. It allows for side-by-side model testing, enabling participating operators to assess performance against their particular languages, regulatory requirements, and market subjects. One of the platform's most compelling features is the ability to create benchmark test sets specifically tailored for low-resource languages and dialects. This includes scenarios that mimic real-world conditions, such as speech-to-text noise, code-switching, and cultural nuances—elements often overlooked in traditional evaluations.

A significant aspect of ModelChorus is the focus on hardware-aware evaluation. This means that not only will operators assess how accurate a model is, but they will also look at latency, memory use, and other efficiency metrics that are crucial for mobile and edge environments. Additionally, the platform supports post-training workflows designed to incorporate feedback and error logs into the training processes, which facilitates ongoing fine-tuning and alignment with target markets.

VEON’s Involvement

VEON Group, a global leader in digital services, is the first mobile operator to join the ModelChorus initiative. By participating from the beginning, VEON brings authentic operator requirements and market conditions to the table, helping to shape the platform's development.

James Kaplan, CEO and Co-Founder of MeetKai, emphasized the significance of ModelChorus, stating, “Our goal is to make language AI measurable, comparable, and deployable in real-world environments, especially for underserved languages and dialects.” This approach signals a commitment to justice in language representation in AI, ensuring that more voices are heard and included in technological advancements.

Further backing the importance of this platform, Louis Powell, GSMA Director of AI Technologies, remarked that as the industry moves toward more robust evaluations of AI performance, ModelChorus will facilitate collaboration between operators and model developers. This union is crucial in enhancing linguistic coverage and improving device performance in mobile settings—critical elements in the pursuit of responsible AI adoption.

The Future of AI Evaluations

Kaan Terzioglu, CEO of VEON, also highlighted that “AI adoption at operator scale requires transparent, repeatable execution, not just demos.” By joining ModelChorus, VEON is contributing to an important evolution in benchmarking AI model readiness, ensuring that the evaluations reflect reality—on real networks, devices, and in the languages that customers engage with daily.

In conclusion, ModelChorus represents a substantial stride toward equitable AI technology in mobile environments. As this frontier in AI continues to evolve, platforms like ModelChorus will be pivotal in guiding the industry toward fair representation across diverse languages and dialects.

For enthusiasts and professionals in the tech industry, the arrival of ModelChorus signifies not only an advancement in AI evaluations but also a hopeful pathway to bridging language divides through technology.