Key facts
- DeepSeek has officially released its V4-Flash AI model, following beta versions of its V4-Pro and V4-Flash models.
- The V4 models offer enhanced knowledge, reasoning, and autonomous agent capabilities.
- DeepSeek's V4 models utilize Huawei chips, reducing reliance on U.S. suppliers.
- The V4-Pro model is priced significantly lower than Western AI offerings.
- DeepSeek claims its V4-Pro Max model surpasses OpenAI's GPT-5.2 and Google's Gemini 3.0-Pro in reasoning benchmarks.
Chinese artificial intelligence startup DeepSeek has released the official version of its DeepSeek-V4-Flash model, featuring enhanced autonomous agent capabilities and significantly reduced API costs. This release intensifies the global AI price war, with DeepSeek's offerings priced substantially lower than U.S. rivals. The V4 models, including V4-Pro and V4-Flash, boast improved knowledge, reasoning, and agentic functions, supported by Huawei chips to lessen dependence on U.S. suppliers. DeepSeek claims its V4-Pro Max model outperforms OpenAI's GPT-5.2 and Google's Gemini 3.0-Pro on reasoning benchmarks, though it falls slightly short of GPT-5.4 and Gemini 3.1-Pro. The V4-Pro model has a 1.6 trillion total parameter count, utilizing a Mixture-of-Experts approach for efficiency, and offers a one million token context window at a claimed 98% lower cost than GPT 5.5 Pro.