The new version, DeepSeek-R1-0528, has a whopping 685 billion parameters, meaning it can perform on par with competitors such as o3 from OpenAI and Gemini 2.5 Pro from Google.

In December, developers observed that DeepSeek’s V3 model often identified itself as ChatGPT, OpenAI’s AI-powered chatbot platform, suggesting that it may’ve been trained on ChatGPT chat logs.

DeepSeek has announced a minor upgrade to its R1 AI model, which previously stunned the tech world by outperforming Western competitors at a fraction of the cost. This news arrives shortly before ...

Chinese startup DeepSeek has released an updated version of its R1 reasoning AI model on the developer platform Hugging Face after announcing it in a WeChat message Wednesday morning. The updated ...

DeepSeek's R1-0528 went live on Hugging Face without any documentation or announcements, but benchmark performance has improved Chinese artificial intelligence (AI) start-up DeepSeek quietly ...

DeepSeek's AI model propelled the Chinese firm to fourth in LiveBench's global LLM rankings amid the US-China AI race. Chinese tech narrows the domestic gap, but monetization barriers persist.

Using a test that evaluates censorship, the developer claimed that the model, DeepSeek R1-0528, is the “most censored” version in terms of responses critical of the Chinese government.

Chinese artificial intelligence (AI) start-up DeepSeek quietly released a new version of its R1 reasoning model on Wednesday, marking its first revision since its high-profile debut in January ...

DeepSeek has released an update to its R1 reasoning model, following Google’s new series of announcements at I/O 2025, as the AI race keeps heating up globally and various tech giants release ...

DeepSeek releases upgraded AI model, claims parity with ChatGPT, Gemini The R1-0528 model features enhanced reasoning capabilities, with DeepSeek asserting it’s closing the gap with OpenAI’s ...

Shanghai-based artificial intelligence (AI) start-up MiniMax has launched an open-source reasoning model that it said requires just half the computing resources of rival DeepSeek’s models for ...