Artificial intelligence (AI) models ChatGPT, Grok, DeepSeek, Claude, o1, and Gemini trade Bitcoin (BTC) and cryptocurrencies and compete in a real-world investment environment, with each allocated USD 10,000.
The experiment, called Alpha Arena, was launched on October 17th and aims to measure the ability of these AIs to operate in open and dynamic markets without human intervention.
According to the latest results, DeepSeek leads the competition with $13,569 total account valuefollowed by Grok at $13,100. Both models show an increase of over 30% compared to the initial capital. In third place is Claude at $12,344, while Gemini remains close to the starting point at $10,717.
Below average were ChatGPT at $7,195 and o1 at $6,715, with both strategies holding short positions and failing to sustain in the face of the market-wide rally recorded on October 19th.
The following graph shows the trading performance of various AIs for Bitcoin, Ether, Solana, BNB, DOGE, and XRP.
Alpha Arena uses real-time market data to execute real trades Same conditions for all models. According to the developers, their goal is to transform artificial intelligence benchmarks into “more realistic” scenarios, where unpredictability and constant competition serve as tests for adaptive intelligence.
Analyst @0xGumshoe commented that the event was “going viral” and highlighted that Grok and DeepSeek had “completed the longest run” and successfully capitalized on Bitcoin's rise.
Other participants, such as user @Cryptomaniac, pointed out that this type of testing could change the way traders choose tools. skill “You will know which AI will trade on your behalf.”
This experiment is currently ongoing and will be one of the first public experiments on artificial intelligence applied to real-world financial transactions.
Last July, CriptoNoticias reported that Generation Z, which includes young people born between 1997 and 2007, is increasingly betting on artificial intelligence as a central tool for decision-making when trading.
