Sorry, OpenAI (and Google and Meta and…).
A recently released AI model called DeepSeek from a China-based startup is currently wreaking havoc on the tech space in the U.S.
Why? Because it's blowing all other Big Tech models out of the water. And, to top it off, it's allegedly doing so with less funding and less technological resources.
Here's what the AI industry says about DeepSeek compared to OpenAI's leading chatbot, ChatGPT.
DeepSeek is actually open
Now, regarding AI outputs, everyone might have a different opinion based on their specific use case. So, there are still areas where other AI models might beat DeepSeek's outputs.
For example, some users discovered that certain answers on DeepSeek's hosted chatbot are censored due to the Chinese government. That is true.
But, here's a fact: DeepSeek is open in a way that OpenAI said ChatGPT would be – and never delivered. In fact, as OpenAI sheds its original "open" ethos, DeepSeek went ahead and released its model as open-source. Anyone can download the DeepSeek R1 model for free and run it locally on their own device. This means your data won't be shared in any way with DeepSeek. In addition, as even DeepSeek pointed out, users can get around any censorship or skewed results.
DeepSeek is more affordable than OpenAI
While OpenAI's training for each model appears to be in multiples of tens of millions of dollars, DeepSeek claims it pulled off training its model for just over $5.5 million.
And that price difference also appears to be passed on to the consumer.
API access for DeepSeek-RI starts at $0.14 for one million tokens or roughly 750,000 words. DeepSeek's latest model is reportedly closest to OpenAI's o1 model, priced at $7.50 per one million tokens. That's a pretty big disparity in pricing.
DeepSeek's outputs for certain tasks seemingly beat ChatGPT
ChatGPT and DeepSeek users agree that OpenAI's chatbot still excels in more conversational or creative output as well as information relating to news and current events.
However, the consensus is that DeepSeek is superior to ChatGPT for more technical tasks. If you use AI chatbots for logical reasoning, coding, or mathematical equations, you might want to try DeepSeek because you might find its outputs better.
For most queries, though, it appears DeepSeek and ChatGPT are on par, roughly giving the same output.
Even being on equal footing is bad news for OpenAI and ChatGPT because DeepSeek is entirely free for most use cases. Regular ChatGPT users may have to subscribe to its paid tier at $20 a month.
For companies utilizing AI-model API access, the price difference between two largely equivalent models may be too much for them not to switch from OpenAI's ChatGPT to DeepSeek.
