February 2026 Summaries
4 posts from Gradium
Filter
Month:
Year:
Post Summaries
Back to Blog
A new partnership between Gradium and InteractionLabs has been announced, with Gradium's advanced voice technology now powering Ongo, a robot designed to enhance emotional presence in home settings. InteractionLabs, led by Creative Director Alec Sokolow, focuses on creating robots with personality, viewing the challenge as both a storytelling and engineering endeavor. The aim is to make robots more relatable and less mechanical by emphasizing voice as a key element of personality and interaction. Gradium's audio language models deliver highly realistic and expressive speech, bridging the gap between human and machine interaction by overcoming current technical limitations in voice AI. This collaboration marks the beginning of a deeper relationship, promising future innovations that will redefine consumer expectations of home robots, with more developments to be announced soon.
Feb 17, 2026
373 words in the original blog post.
Gradium specializes in creating advanced audio language models that facilitate natural and expressive voice interactions with ultra-low latency, suitable for a range of voice tasks including voice cloning and conversational AI. These models are accessible via API and compatible with various NVIDIA GPUs, ensuring flexible deployment to meet different performance and scalability requirements. The models focus on key performance metrics such as Time To First Audio (TTFA) and Real-time Factor (RTF), crucial for interactive voice AI applications to maintain responsiveness and avoid audio skips. Utilizing Delayed Streams Modeling (DSM) architecture, Gradium's models allow for efficient batching and streaming, employing techniques like the CUDA Graph API to minimize latency. The models are optimized for NVIDIA hardware, balancing audio quality, latency, and GPU costs through adjustable codebooks. These optimizations enhance the performance of voice AI agents, offering seamless and responsive conversational experiences that improve user satisfaction and engagement, ultimately providing a competitive edge for businesses deploying voice AI solutions.
Feb 11, 2026
1,309 words in the original blog post.
Gradium is developing advanced real-time voice models, including speech-to-text (STT) and text-to-speech (TTS) technologies, with features like voice cloning and semantic turn-taking, aimed at empowering startups in various industries such as gaming and customer care. To address the financial and infrastructural challenges that startups face, especially concerning high voice API pricing, Gradium has launched the Gradium Startup Program, offering six months of their M Plan free to eligible startups. This program includes substantial benefits like monthly free credits, high-fidelity voice cloning, and commercial rights, designed to help startups with fewer than 25 employees who have secured initial funding and are focused on voice-first applications. Participants also gain access to Gradium's engineering support and early previews of new research, fostering innovation and allowing startups to create unique audio identities while scaling their operations.
Feb 09, 2026
545 words in the original blog post.
Acolad, a leader in language and content solutions, and Gradium, specializing in AI models for voice interactions, have announced a strategic partnership to develop AI-powered interpreting solutions for enterprise and public-sector environments. This collaboration aims to address challenges such as data quality, multilingual coverage, and cultural nuance in AI adoption, by integrating Acolad's language expertise and AI-driven workflows with Gradium's advanced voice AI models. The partnership is designed to create scalable, reliable, and accurate real-time multilingual communication tools, leveraging Acolad's extensive global experience and Gradium's cutting-edge voice synthesis and transcription technologies. Both companies are committed to setting a new standard for AI-powered multilingual communication that is ready for real-world deployment, ensuring quality and operational resilience with human expertise as needed. This venture not only highlights Acolad's role in training and governing AI voice technologies but also demonstrates Gradium's ability to translate research excellence into impactful real-world applications.
Feb 05, 2026
558 words in the original blog post.