Archive — CN — technode
Ecosia switches from Mistral to open-weight AI models including Qwen, GLM andKimi
Ecosia, the German search engine, has switched from Mistral to open-weight AI models including Qwen, GLM, and Kimi, via a partnership with Melious, reducing costs by half while improving performance. Founder Christian Kroll cited issues with Mistral such as overloaded servers and unsatisfactory model quality, prompting the change.
Key points
- Ecosia is adopting open-weight AI models to enhance its search capabilities.
- The switch to Qwen, GLM, and Kimi is facilitated by collaboration with Melious.
- Switching from Mistral has led to cost reduction by approximately 50%.
- Ecosia faced technical issues with Mistral, leading to dissatisfaction with its performance.
- The partnership allows access to multiple AI models, improving flexibility and reliability.
Why it matters
- The move reflects a shift in the AI landscape towards open-source models, which is a growing trend in 2025 IT.
- Many organizations are moving away from single-vendor solutions to reduce risks and costs.
- Ecosia's decision aligns with broader trends towards sustainability and efficiency in tech.
Ecosia's transition to open-weight AI models highlights a significant trend in tech towards flexibility and cost efficiency, influencing how AI is integrated into services while addressing past challenges.