DeepSeek V4 Launching Next Week with Image and Video Generation
Original: DeepSeek V4 will be released next week and will have image and video generation capabilities, according to the Financial Times View original →
DeepSeek V4 Imminent Release
According to an exclusive Financial Times report, Chinese AI startup DeepSeek is set to release V4 next week, marking another significant leap in the company's rapidly advancing capabilities.
Multimodal Capabilities: Image and Video Generation
The most notable new feature is integrated image and video generation. This represents DeepSeek's transition from a text-focused LLM to a full multimodal AI platform, placing it in direct competition with GPT-4o and Gemini's multimodal capabilities.
New Challenge to US Rivals
DeepSeek already disrupted the AI industry with V3 and R1 last year, demonstrating that frontier-level performance could be achieved at a fraction of the cost. V4's launch continues this trajectory, particularly remarkable given ongoing US restrictions on advanced semiconductor exports to China.
Open-Source Policy
Whether DeepSeek will maintain its open-source approach with V4 is a key question for the developer community. Previous releases under open licenses had significant downstream impact on the global LLM ecosystem, accelerating local deployment and fine-tuning efforts worldwide.
Related Articles
Anthropic is not only shipping a stronger Claude model; it is splitting the same base capability into a broad Fable release and a restricted Mythos track. The package includes $10/$50 token pricing, 30-day safety retention, and automatic fallback to Opus 4.8 for some high-risk requests.
Anthropic has accused Chinese AI firms of creating over 24,000 fraudulent accounts to extract 16 million training exchanges from Claude for model distillation.
HN did not latch onto DeepSeek V4 because of a polished launch page. The thread took off when commenters realized the front-page link was just updated docs while the weights and base models were already live for inspection.