AI BriefingDeepSeek v4 Released
·2026.04.24 12:14
Key point
DeepSeek has unveiled two variant models, **V4-Pro** and **V4-Flash**.
Details
DeepSeek has unveiled DeepSeek-V4-Pro and DeepSeek-V4-Flash.
- DeepSeek-V4-Pro: 1.6T parameters, 49B active parameters
- DeepSeek-V4-Flash: 284B parameters, 13B active parameters
It appears to have been introduced via Hugging Face, and the key point is that a new version of the large-scale model family and a lightweight variant were released together.
This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.
Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.
Related
Toss Doctor V2 Released: Maximizing QA Efficiency with LLM-based Automated Testing and Self-healing