share_log

DeepSeek-V3模型更新,各项能力全面进阶

The DeepSeek-V3 model has been updated, with comprehensive enhancements across all abilities.

Breakings ·  Mar 25 13:45

The DeepSeek-V3 model has completed a minor version upgrade, and the current version number is DeepSeek-V3-0324. The new V3 model leverages reinforcement learning techniques used during the training of the DeepSeek-R1 model, significantly improving performance levels in inference tasks, achieving scores that exceed GPT-4.5 in mathematical and coding-related evaluation benchmarks. The new V3 model has also seen a considerable enhancement in capabilities in aspects such as tool usage, role-playing, and casual Q&A. (Securities Times)

The translation is provided by third-party software.


The above content is for informational or educational purposes only and does not constitute any investment advice related to Futu. Although we strive to ensure the truthfulness, accuracy, and originality of all such content, we cannot guarantee it.
72
Comment Comment 20 · Views 156.8k

Comment(20)

Recommended

Write a comment
20

Statement

This page is machine-translated. Futubull tries to improve but does not guarantee the accuracy and reliability of the translation, and will not be liable for any loss or damage caused by any inaccuracy or omission of the translation.