4 hours Ago
DeepSeek R2 AI model launch is expected soon and reports say it will be even cheaper to make than the ChatGPT 4o model. ChatGPT’s rival from China, DeepSeek is widely expected to release the new R2 AI model sometime this month. According to reports from China, the new DeepSeek AI model is going to be cheaper and much better, making the more expensive ChatGPT 4o look like a costly affair.
We have already seen the abilities of the R1 and V3 models, and it seems the R2 version is likely to take reasoning to another level. The ongoing US-China tariff war is also going to be in the spotlight as the latter looks to build its own AI empire. DeepSeek R2 AI: Cheaper And Better? Reports from China claim the R2 AI model will be developed using a mixed architecture which is allowing DeepSeek to create the model at 97 percent less cost than what OpenAI spends on ChatGPT 4o and its features like Ghibli style images.
The use of the technology helps DeepSeek reduce its computing expense which is the biggest outlay for most of the AI companies these days. The new R2 model promises to be improved in reasoning, better with coding and perform much higher in the benchmark ranks compared to the US-based rivals. It is also being said that unlike most companies, DeepSeek is going to depend on local sources like Huawei for the GPUs which keep the data systems running and let the AI do its work efficiently.
DeepSeek has raised concerns outside of China, purely because the company’s policy says the data is shared with servers in the home region. The Indian government has understandably blocked use of ChatGPT and DeepSeek R1 models for its employees. Apple is heavily relying on local developers to get its AI features rolling out in China and recently, Apple chief Tim Cook was asked about the DeepSeek AI company by the state media during his visit, to which he said “excellent.
" That’s not all, the Apple chief reportedly said that he recommends AI models of Chinese startups like DeepSeek..
