Mistral is launching a 1-trillion-parameter open AI model it claims leads outside China
The French startup says ML4, trained on 4,000 Nvidia GPUs in Europe, leads open-weight models outside China by a substantial margin
On September 22, a flurry of artificial intelligence (AI) model releases swept the global tech landscape. Xiaomi, based in Beijing, unveiled its MiMo-V2.6 model during a live stream, while AI pioneers Anthropic, based in San Francisco, introduced the Opus 5.5 model. Shortly after, OpenAI dropped the surprise releases of its GPT-6 Sol and Luna models.
This rapid fire of AI model launches presented tech executives and developers with an overwhelming challenge. However, while Silicon Valley grappled with what was dubbed "model fatigue," China's tech sector faced the phenomenon on an amplified scale. A one-day release cycle, following major roll-outs from competitors such as Z.ai, DeepSeek, Tencent Holdings, Alibaba Group Holding, and Moonshot AI, demonstrated how Chinese companies were rapidly churning out upgrades, rendering individual breakthroughs difficult to stand out.
Rui Ma, a tech analyst and founder of the Tech Buzz China newsletter, suggested that model fatigue might be more pronounced in China. The intense pace of releases has led to brutal price wars, with each new generation of models resetting the price floor for customers. Chinese users are generally more price-sensitive, making brand loyalty challenging to establish at the model level.
Analyst Lionel Sim argued that the economic pressure to adopt newer models in China was particularly heavy, as each upgrade reset the price floor for customers. Companies like DeepSeek, Z.ai, Alibaba, and Xiaomi adopted various strategies to cope with the pricing squeeze, such as slashing API prices, claiming cost-effective performance, or focusing on higher intelligence without increasing costs.
US developers, such as Anthropic and OpenAI, also adjusted to the economic pressures by reducing operating costs and halving API prices for newer models. Despite the overwhelming pace of releases, AI demand remains robust, with enterprise clients prioritizing stability, predictable output, integration costs, data security, and technical support over minor benchmark improvements.
To survive the churn, Chinese tech firms focus on specialisation rather than competing solely on general benchmark scores.
Written by urgent.news from South China Morning Post's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
This story
This is one outlet's version. Read the fullest account.
