Solaria delivers unmatched accuracy, speed, and native-level transcription in 100 languages—including 42 underserved by any other STT model.
Gladia, an AI transcription and audio intelligence provider, announced the launch of Solaria, a next-gen automatic speech recognition (ASR) model designed to redefine real-time communications for call centers and other voice-first platforms.
While outsourcing has long been a cost reduction strategy in the call center industry, businesses now face a new, critical challenge: providing seamless, multilingual support at scale. With 49% of global executives reporting financial losses due to language barriers1, the demand for scalable, high-quality multilingual solutions has never been greater. Solaria now empowers businesses to enhance and expand their customer service operations with AI-powered voice technology that delivers unmatched language coverage—supporting 40+ languages previously inaccessible with other solutions—without compromising quality or speed.
Solaria: An Enterprise-Ready Model for Global Customer Experience
The only speech-to-text (STT) engine built for true global scalability, Solaria was designed to meet the demands of today’s contact centers, where both AI automation and human agents need high-accuracy, low-latency, and real-time multilingual support to succeed.
The model achieves industry-leading results in speech recognition, delivering both unmatched accuracy and ultra-fast processing speed. Recent benchmarks show Solaria has reached an unmatched 94% Word Accuracy Rate (WAR) average in English, Spanish, French and other common languages, while maintaining an ultra-low latency of 270 ms, making the conversation feel natural and responsive.
While real-time STT is often measured by speed alone, accuracy and language coverage are equally crucial for businesses providing seamless services across regions.
Also Read: TransPerfect’s GlobalLink named leader in G2 Spring 2025 for Enterprise Translation Management
Unlike other STT models that prioritize speed over usability, Solaria balances industry-leading accuracy and speed with unmatched language coverage—100 languages in total, with exclusive support for 42 languages not matched by competitors. For high-population markets and key outsourcing hubs like Bangladesh, India, and The Philippines, native-level accuracy in regional languages is now offered through Solaria.
With native-level transcription, real-time code-switching, and translation across all supported languages, businesses can expand into global markets without constraints.
Designed for enterprise-scale voice automation, Solaria delivers:
- Best-in-class accuracy in high-population languages such as Tagalog, Bengali, Punjabi, Tamil, Urdu, Persian, and Marathi.
- Ability to adapt the model to industry-specific terminology (like medical or financial jargon) and have it extract critical data, like names, addresses, and numericals.
- Adaptive speech processing, ensuring high accuracy in noisy call center environments.
- Enterprise-grade data security, in full compliance with GDRP, HIPPA, and SOC 2.
With the addition of Solaria to its product portfolio, Gladia allows businesses to enhance customer service by improving AI-powered voice agents, making IVRs and virtual assistants more reliable across multiple languages, while also optimizing human-assisted workflows with real-time transcriptions and translations to help agents provide more effective assistance.
“Speech is the most natural way to connect with the world—for the first time, automated speech recognition is closing the divide, enabling humans and AI to truly speak the same language,” said Jean-Louis Quéguiner, Chief Executive Officer and Co-Founder, Gladia. “With Solaria, we have made a breakthrough in AI-powered voice technology that unlocks new opportunities for businesses, driving efficiency and delivering more seamless, impactful customer experiences across diverse languages and markets. Solaria is built for next-generation voice platforms ready to lead this transformation on a global scale.”
SOURCE: PRNewswire