Faster, smarter DeepSeek V3.1 AI arrives for free

[Redaktion]

DeepSeek-V3.1 AI LLM has arrived eight months after the initial launch of V3. The chatbot software is able to answer prompts quicker and smarter than before. The model is open source, allowing anyone to use and build upon it for free.

https://www.notebookcheck.net/Faster-smarter-DeepSeek-V3-1-AI-arrives-for-free.1097372.0.html

[Alfar]

Interesting, but the token limit is not impressive. Complex problems require very high token counts. I don't get to choose my LLM, simply because there's only two that are capable of non-toy problems.

I've no problems with others using the problem sets I've devised to test LLMs, and I did use DeepSeek - it is very good - when the problem sets were small enough, but the token count needs work.