Today for AI

IT之家 · 智能时代 · 10/8/2026, 15:00:05

Ecosia Dumps Mistral for Chinese Open-Source Models, Halving Costs

By IT之家Original title: 不满 Mistral 表现,德国搜索引擎 Ecosia 押注中国开源 AI 模型
68AI Score
Executive Summary

German search engine Ecosia has terminated its partnership with Mistral due to concerns over quality lag, server instability, and environmental compliance, switching instead to Chinese open-source models like Qwen, GLM, and Kimi. This strategic pivot halved Ecosia's AI service costs while improving performance, highlighting the growing competitiveness of Chinese open-source models in European commercial applications.

SOURCE COVERAGEOriginal coverage

Dissatisfied with Mistral's Performance, German Search Engine Ecosia Bets on Chinese Open-Source AI Models

In an interview, Ecosia founder and CEO Christian Kroll stated: "We are disappointed with the quality of Mistral; the model is now a full year behind its competitors."

This past May, Ecosia announced it was dropping OpenAI’s solutions in favor of switching its AI service provider to Mistral. According to Kroll, the partnership has not been smooth, with the company frequently encountering technical issues such as server overloads.

As reported by IT Home, Ecosia also has concerns regarding whether Mistral aligns with its environmental goals. France’s energy mix relies heavily on nuclear power, and the company questions whether Mistral can maintain neutrality. Kroll explained: "Mistral depends on funding from international investors, so it cannot be considered a truly sovereign enterprise."

The CEO further believes that China’s active promotion of open-source large language models represents a significant opportunity for Europe. Although Europe missed the first phase of AI model development, it can now leverage increasingly powerful open-source models without investing billions of dollars to train frontier AI systems independently.

Consequently, Ecosia has chosen to partner with Melious, utilizing China’s Qwen, GLM, and Kimi models. This shift has halved their costs while delivering performance improvements.