In a significant update, Chinese AI startup DeepSeek has launched the latest DeepSeek-V3.1.
It's a major upgrade to its flagship model that improves processing speeds and supports the high-end Chinese-manufactured semiconductor chips.
This move from the Chinese based tech giant highlights the country’s push towards excellence in the technological landscape amid US export restrictions on cutting-edge chips.
Taking to WeChat, Chinese based instant messaging app, DeepSeek stated that the recently launched V3.1 model integrates the UE8M0 FP8 precision format, particularly designed for efficiency with forthcoming domestic processors.
FP8 enables models to use less memory while running faster, making it a substantial choice for large-scale AI applications where cost and speed are essential.
While the company has yet to disclose the chipmaker's names.
Additionally, the model rolls out a hybrid inference structure, allowing users to switch between reasoning and non-reasoning models.
A “deep thinking” button in DeepSeek’s app and web platform will allow users to activate more advanced processing for complex queries.
DeepSeek further revealed the revised API pricing will begin from September 6, aiming to keep its company affordable.
The company garnered significant attention for offering a range of AI models, aiming to outdo third-party tools such as ChatGPT at lower costs.
