DeepSeek Unveils V4 AI Model with Enhanced Capabilities and Huawei Integration
Solo los hechos

DeepSeek Unveils V4 AI Model with Enhanced Capabilities and Huawei Integration

Resumen

Chinese AI startup DeepSeek has released preview versions of its V4 model, featuring significant improvements in knowledge, reasoning, and autonomous task performance, while integrating Huawei's chips to reduce reliance on U.S. hardware.

Chinese artificial intelligence startup DeepSeek has introduced preview versions of its latest AI model, V4, marking a significant advancement in the ongoing technological competition between China and the United States. The V4 series includes two open-source models, "Pro" and "Flash," both designed to enhance knowledge, reasoning, and autonomous task execution capabilities. Notably, these models are optimized to operate on Huawei's Ascend AI processors, reducing DeepSeek's dependence on U.S. hardware suppliers such as Nvidia.

The V4-Pro model boasts 1.6 trillion parameters, while the V4-Flash model features 284 billion parameters. Both models support a context window of one million tokens, allowing them to process and recall extensive information.

DeepSeek's previous model, R1, released in January 2025, garnered global attention for its cost-effective reasoning capabilities, signaling China's progress in AI development.

The release of V4 comes amid allegations from U.S. AI firms OpenAI and Anthropic, accusing DeepSeek of illicitly extracting capabilities from their models to enhance its own.

Industry analysts view the V4 rollout as a pivotal milestone for China's AI industry, especially as global competition intensifies in the pursuit of self-reliance in critical technologies.

Fuente

AP News

Sigue leyendo, gratis

Crea una cuenta gratis para seguir las noticias. No se requiere tarjeta.