Deepseek R2 must be released before planning, newspaper found

Deepseek wants to start the R2 model as soon as possible, Reuters. The Chinese company plans to take advantage of the impulse and the fame conquered by the launch of artificial intelligence R1.

Initially, Deepseek planned to launch R2 at the beginning of May 2025, but the sources heard by the report Company plans to advance the program. However, the new launch window has not been announced.

Very little is known about Deepseek Ai R2, However, it is likely that the model is a direct evolution of R1.With a wider context window for coding, reasoning and requests. It can also provide extended support for more languages.

. Deepseek R2 can also earn multi -mod featuresIncluding tools for interpretation and creation of images and videos based on text descriptions.

The R1 carried a kilometer for Deepseek, came to the fore for the efficiency of the model, and reduced training costs compared to Openai GPT-4. The launch had a significant impact on the market, overthrowing Nvidia’s actions, causing Openai, who accused the Chinese company of distillation to educate Chatgpt to educate its own model.

Until now, Deepseek did not comment on the rumor.

Did you know that Deepseek censored the answers when asked about the problems of the Sensitive Chinese government? . Tecmundo He activated these restrictions and saw how the R1 outputs were different compared to competitors.