GMAsia
    🇨🇳China·Startups·9 Oct 2026·via Tech Node

    Ecosia switches from Mistral to open-weight AI models including Qwen, GLM and Kimi

    German search engine Ecosia has replaced Mistral with open-weight AI models, including Chinese-developed Qwen, GLM, and Kimi, through a partnership with German AI platform Melious. Founder and CEO Christian Kroll stated that the change roughly halved costs while improving quality and performance, according to the report.

    Nexa's Summary

    Ecosia's decision to switch from Mistral to a multi-model strategy, incorporating Chinese-developed open-weight models, reflects a practical response to operational challenges. The company previously adopted Mistral in May after moving away from OpenAI, but reported encountering recurring technical problems, such as overloaded servers, and dissatisfaction with model quality. This suggests that even established AI providers may face scalability and performance hurdles that can drive users to seek alternatives.

    The reported cost reduction, approximately halving expenses while simultaneously improving quality and performance, highlights a significant benefit of Ecosia's new approach. By accessing multiple models through Melious, Ecosia avoids reliance on a single AI developer. This diversification offers flexibility, allowing the search engine to select models best suited for specific tasks or to switch providers if performance or cost issues arise with one model.

    The inclusion of Chinese-developed models like Qwen, GLM, and Kimi in Ecosia's new lineup demonstrates their growing competitiveness and viability in the global AI market. For companies seeking cost-effective and high-performing solutions, these models represent alternatives to Western-developed offerings. This trend could encourage further adoption of diverse AI models across various applications, driven by practical considerations of cost, quality, and technical reliability.

    Share this article

    Go deeper
    Original reporting by Tech NodeWe don't republish, read the full story â†’

    Related reading

    6 stories