China's DeepSeek previews new model adapted for Huawei chip technology
The Chinese startup says the open-source models rival top closed systems while cutting token costs to as low as $0.14 per million input tokens.
- The Hangzhou-based DeepSeek released a "preview" version of its DeepSeek-V4 model on Friday, aiming to compete with top-tier artificial intelligence systems from American rivals like OpenAI and Google DeepMind.
- Founded in 2023, the startup gained global attention with free, open-source models that historically achieved performance benchmarks matching leading LLMs at a fraction of the cost of U.S. competitors.
- The new DeepSeek-V4 "features an ultra-long context of 1 million words, achieving leadership in both domestic and open-source fields," the company said; it is available in DeepSeek-V4-Pro and DeepSeek-V4-Flash versions.
- Optimized for agent tools like Anthropic's Claude Code and OpenClaw, DeepSeek-V4 trails Google's Gemini-Pro-3 in world knowledge benchmarks, though the company claimed strong results against domestic competitors.
- Amid intensifying technological competition, reports suggest DeepSeek optimized the model for Huawei Technologies' Ascend chips, while officials prior to release accused the company of using banned Nvidia Blackwell hardware.
168 Articles
168 Articles
China’s DeepSeek rolls out a long-anticipated update of its AI model
HONG KONG (AP) — DeepSeek, the Chinese artificial intelligence startup that shook world markets last year, launched preview versions of its latest major update Friday as the AI rivalry between
The Chinese startup that surprised the world with its levels of development last year showed a great update, with the advantage of no longer relying on chips made by US companies
Coverage Details
Bias Distribution
- 50% of the sources are Center
Factuality
To view factuality data please Upgrade to Premium


























