Deepseek surpasses ChatGPT to top the AppStore. What are the related AI concept coins?

avatar
Wenser
1 days ago
This article is approximately 1184 words,and reading the entire article takes about 2 minutes
Technology and Meme are intertwined. Can Deepseek lead the next wave of AI Agent craze?

Original | Odaily Planet Daily ( @OdailyChina )

Author: Wenser ( @wenser 2010 )

Deepseek surpasses ChatGPT to top the AppStore. What are the related AI concept coins?

After Xiaohongshu topped the US AppStore due to the Tiktok ban, a Chinese dark horse product once again topped the AppStore in the AI field.

What is quite different is that the value of this ranking is even greater than that of Xiaohongshu. According to media reports , DeepSeek topped the free APP download rankings of Apples App Store in China and the United States, and surpassed ChatGPT, a benchmark product in the field of AI, in the US download rankings. At the same time, the concept tokens related to the Deepseek R 1 model have also been popular due to market attention, and the market value of many tokens has exceeded the threshold of 10 million US dollars.

Odaily Planet Daily will briefly introduce Deepseek products and related concept tokens in this article for readers reference.

Deepseek: A tough challenge to OpenAI, a better and cheaper R1 model than ChatGPT o 1

The origin of all the heat lies in an article officially released by Deepseek 7 days ago.

On January 20, an article titled DeepSeek-R 1 released, performance benchmarked against OpenAI o 1 official version was published, directly challenging the o 1 model of AI giant OpenAI. It gave a detailed introduction to Deepseeks R 1 model, mentioning:

DeepSeek-R 1 uses reinforcement learning technology on a large scale in the post-training stage, which greatly improves the models reasoning ability with only a small amount of labeled data. In tasks such as mathematics, code, and natural language reasoning, its performance is comparable to the official version of OpenAI o 1.

We have made all the DeepSeek-R 1 training technology public in order to promote full communication and innovative collaboration in the technical community. Paper link: https://github.com/deepseek-ai/DeepSeek-R1/blob/main/DeepSeek_R1.pdf

The DeepSeek-R1 API service is priced at 1 yuan (cache hit) / 4 yuan (cache miss) per million input tokens and 16 yuan per million output tokens.

Simply put, the R1 model launched by Deepseek not only has performance comparable to ChatGPT o 1, but also directly open-sources the relevant training technology. At the same time, its API pricing is even shockingly low.

When many people were still wondering whether Deepseek was boasting, a series of verification results from industry insiders later confirmed that what Deepseek said was true! And the sharp product that can cover both B-end and C-end related needs has also made countless American AI companies and teams exclaim: The wolf is coming!

Jiayi Pan, a PhD candidate at the University of California, Berkeley, and his team have successfully reproduced the key technology aha moment in DeepSeek R 1-Zero at a very low cost. The research shows that even with smaller language models, through reinforcement learning (RL), the model can develop strong self-verification and search capabilities on its own. It is worth noting that the research team showed that the cost of training the model is less than $30. The above-mentioned aha moment is the key technology in DeepSeek R 1-Zero.

According to the introduction of the technology in the DeepSeek-R 1 technical report, through the RL framework, AI may spontaneously develop human-like reasoning capabilities and even surpass the limitations of preset rules. And this will also hopefully provide direction for the development of more autonomous and adaptive AI models, such as dynamically adjusting strategies in complex decision-making (medical diagnosis, algorithm design).

Previously, a Meta employee posted on TeamBlind, an anonymous workplace community in the United States, that due to the excellent performance of DeepSeeks models and extremely low training costs, Metas engineers are frantically analyzing DeepSeek and trying to copy some of its content.

In addition, AI blogger scaling 01 conducted a comparative analysis and found that DeepSeek not only surpassed in performance, but also had a training cost of only $5.58 million for DeepSeek-V3, which has 671 B parameters. In comparison, the computational cost of Metas Llama 3 series models is enough to train DeepSeek-V3 at least 15 times.

——Quoted from APPSO related articles

It is worth mentioning that Deepseek is backed by Huanfang, a well-known domestic quantitative fund. It had previously stockpiled a large number of high-performance graphics cards due to the needs of quantitative trading. And because it is a financial company, it has not been included in the US sanctions list of technology companies. Ultimately, relying on its extremely strong determination and long-term technical route, it finally achieved the current result of using smaller data volumes to achieve better output models.

Of course, Meme coin, which has always been a barometer of news hotspots, is still ahead of the market.

Deepseek-related concept tokens: market hotspots where technology and Meme are intertwined

The following are Deepseek-related concept tokens that are currently receiving some attention in the market.

Seek: AI concept coin with the same name, market value reported at US$22.7 million

As early as December last year, when Deepseek V3 was released, many people noticed its hype value. Seek was the first AI Meme coin of the same name that borrowed the concept of Deepseek. It was born 20 days ago.

According to GMGN data , its current price is $0.022, and its market value is $20.6 million. Previously, its market value exceeded $33 million, and its 24-hour trading volume was about $54.6 million. It is worth mentioning that Deepseek’s official X platform account had previously stated that Deepseek had not officially released any tokens.

Deepseek surpasses ChatGPT to top the AppStore. What are the related AI concept coins?

seek Token Information

POLAR: AI Copilot built on Deepseek, with a market value of US$11 million

Polar’s main concept is “the first AI copilot in the Solana ecosystem, based on a framework built on Deepseek.” According to its latest tweet , it has achieved a series of milestones in just three days since its launch, and the Polar API is about to be opened.

According to GMGN data , its current price is $0.011 and its market value is around $11 million. Previously, its market value once exceeded $22 million, and its 24-hour trading volume was approximately $36.4 million.

Deepseek surpasses ChatGPT to top the AppStore. What are the related AI concept coins?

POLAR Token Information

AAA: Former Amazon programmer issues AAA token, market value reported at $6.1 million

AAA is an AI Agent- related concept token released by a former Amazon programmer named nim. According to his tweets , it uses the Deepseek model and its cost is reduced by 98% compared to the ChatGPT o 1 model.

According to GMGN data , its current price is $0.0061, and its market value is around $6.1 million. Previously, its market value exceeded $7.5 million, and its 24-hour trading volume was approximately $4.5 million.

Deepseek surpasses ChatGPT to top the AppStore. What are the related AI concept coins?

AAA Token Information

The latest news is that ai16z founder Shaw posted: By the way, Eliza has completed the support for DeepSeek two weeks ago. It has to be said that ai16z is still at the forefront of tracking the AI model track. In the future, we will continue to pay attention to more AI concept-related tokens.

In addition, Odaily kindly reminds you: The market price of Meme coin fluctuates greatly. The above content does not constitute investment advice. Please participate with caution, DYOR.

Original article, author:Wenser。Reprint/Content Collaboration/For Reporting, Please Contact report@odaily.email;Illegal reprinting must be punished by law.

ODAILY reminds readers to establish correct monetary and investment concepts, rationally view blockchain, and effectively improve risk awareness; We can actively report and report any illegal or criminal clues discovered to relevant departments.

Recommended Reading
Editor’s Picks