With the innovative release of its most recent large-scale language model (LLM), which is on track with top models from industry titans like OpenAI, Chinese firm DeepSeek has greatly impacted the generative AI scene. According to the Hangzhou-based company, DeepSeek's V3 model, which was trained in just two months with much less computer power, performs on par with the best proprietary model in the world, GPT-4o, at a substantially lower cost than its competitors.
With the innovative release of its most recent large-scale language model (LLM), which is on track with top models from industry titans like OpenAI, Chinese firm DeepSeek has greatly impacted the generative AI scene. According to the Hangzhou-based company, DeepSeek's V3 model, which was trained in just two months with much less computer power, performs on par with the best proprietary model in the world, GPT-4o, at a substantially lower cost than its competitors.