Orange has announced an expansion of its partnership with OpenAI to deploy two advanced open-weight reasoning models, ‘gpt-oss-120b’ and ‘gpt-oss-20b’, within its data centre infrastructure. This move aims to enhance AI solution development support across Orange’s global footprint, with a particular focus on Africa.
The new models will be flexibly deployed across Orange’s data centres in 26 countries, enabling local hosting of AI workloads. This approach safeguards sensitive data while ensuring compliance with varied national regulations across Europe, the Middle East, and Africa.
A key focus is accelerating the integration of regional African languages into large language models (LLMs). By fine-tuning OpenAI’s gpt-oss models with local languages, Orange will empower customers in its 18-country African footprint to interact naturally with AI-powered customer support, sales, and marketing tools.
Orange also plans to release these customized AI models as open-source resources, freely available to local governments for public service applications.
Deploying OpenAI’s reasoning models within Orange’s infrastructure will foster innovative applications, including natural-language voice assistants, chatbots, smarter network automation, enhanced voice recognition services, and AI-driven analysis of sensitive data tailored to specific client needs.
Steve Jarrett, Orange’s Chief AI Officer, stated, “This strategy drives new use cases to address sensitive enterprise needs, help manage our networks, enable innovative customer care solutions including African regional languages, and much more.”
