
DeepSeek takes major step in AI race, beats Alibaba and challenges OpenAI
DeepSeek launched a new AI model, R1-0528, which outperforms Alibaba’s Qwen3, matches OpenAI, and is gaining wide adoption in China and globally.

DeepSeek launched a new AI model, R1-0528, which outperforms Alibaba’s Qwen3, matches OpenAI, and is gaining wide adoption in China and globally.

DeepSeek shares an upgraded version of its R1 AI model on Hugging Face, now available under MIT licence with 685 billion parameters.

President Brad Smith says Microsoft bans staff from using the DeepSeek app due to data security risks and fears of Chinese propaganda.

Microsoft launches compact Phi 4 AI models that offer powerful reasoning skills on smaller devices, rivaling much larger systems.

DeepSeek quietly releases Prover-V2, a maths-focused AI model, just after Alibaba’s Qwen3 launch, boosting speculation in the AI race.

Meta releases Llama 4 AI models, offering faster, smarter tools with new licensing rules and broader answers on sensitive topics.

Ant Group says it cut AI training costs by 20% using Alibaba and Huawei chips, challenging Nvidia’s dominance in the AI chip market.

Baidu launches Ernie 4.5 and Ernie X1, claiming they surpass DeepSeek and OpenAI in AI benchmarks while shifting towards open-source AI development.

OpenAI calls DeepSeek “state-controlled” and urges bans on PRC-backed AI models, citing security concerns and risks of data sharing under Chinese law.