# DeepSeek rolled out preview versions of a new flagship artificial intelligence model.

*genai · news · 2026-04-24 · The Hindu Business Line*

## Key points

- DeepSeek has unveiled its new flagship AI models, V4 Flash and V4 Pro, in preview.
- The V4 models can process a maximum output of 384,000 tokens per request.
- DeepSeek's 2025 release of R1 triggered a trillion-dollar stock market selloff.
- American tech giants are projected to spend $650 billion on AI infrastructure in 2026.

**Companies:** DeepSeek
**Countries:** China

[Read the full story on The Hindu Business Line](https://www.thehindubusinessline.com/info-tech/deepseek-unveils-new-flagship-ai-model-a-year-after-breakthrough/article70900293.ece)

---

Canonical: https://newsio.io/n/81c447df-b477-4786-a952-f82231478f6e/deepseek-rolled-out-preview-versions-of-a-new-flagship-artificial-intelligence-m
Summarized by Newsio from The Hindu Business Line. https://newsio.io/how-it-works