DeepSeek V4: Affordable AI Powerhouse

DeepSeek unveils V4: A game-changer in AI models

DeepSeek, a Chinese AI lab, has released preview versions of their latest AI model, DeepSeek V4, with Flash and Pro editions. These models boast impressive features such as one million token context windows and a staggering 1.6 trillion total parameters, making them the largest open-weight models available. The Pro version outperforms rivals like Moonshots Kimi K two point six.

DeepSeek V4 promises improved efficiency and performance across reasoning benchmarks, nearly matching top closed models. The Pro version even edges out OpenAIs GPT-five point two or Gemini three point zero Pro on key tasks.

The pricing of DeepSeek V4 is competitive, with V4 Flash running at fourteen cents per million input tokens and twenty-eight cents output, undercutting GPT-five point four Nano and others. The Pro version is even cheaper, at fourteen point five cents input and three dollars forty-eight cents output.

Despite their impressive capabilities, DeepSeek V4 models lag slightly on knowledge tests behind GPT-five point four and Gemini three point one Pro, trailing frontier leaders by three to six months in development. The release of these models comes amidst claims of China swiping American AI IP, with DeepSeek facing allegations of distilling tech from Anthropic and OpenAI.

The global AI arms race is heating up, with these affordable powerhouses set to shift the field.

Support the show:
Get a discount at https://solipillow.com/discount/dnn.

Advertise on DNN:
advertise@thednn.ai

This is an automated, high-level news summary based on public reporting.
Report issues to feedback@thednn.ai.

View sources & latest updates:
https://sources.thednn.ai/180cc5597a27c60a