Победитель первого дня турнира Alpha Arena — модель Gemini-3-Pro с рекордным профитом Translation: Headline: Day One Winner of Alpha Arena Tournament — Gemini-3-Pro Model with Record Profit

A new round of the Alpha Arena trading tournament for AI bots, organized by the nof1 laboratory, has commenced. On the opening day, the Gemini-3-Pro model surged to the lead with an average profit exceeding $500.

Similar to the first phase, the neural networks were given $10,000 and the autonomy to make trading decisions. Their goal is to out-earn their competitors (or incur lesser losses).

In season 1.5, the artificial intelligence from various companies is focusing solely on U.S. stocks. In addition to Qwen3, DeepSeek, Claude Sonnet, Gemini, Grok, and GPT 5, two new models have been introduced: Kimi 2 and a “secret LLM from a leading AI lab.”

In this new competition, the models will participate in multiple challenges, each with its distinct theme and specific parameters.

The progress of results can be monitored across each topic. For instance, by utilizing maximum leverage, Gemini has already reached $11,743 from the initial $10,000, whereas DeepSeek has unfortunately dropped by over $1,000.

The Aggregate Index displays the average trading performance across all four themes.

Overall, various configurations will stress-test the neural networks and demonstrate their resilience in diverse scenarios.

We recall that on November 3, the first season of Alpha Arena concluded, with four of the six participating models incurring losses.

In October, former Binance CEO Changpeng Zhao predicted a rise in the popularity of AI-assisted trading following the success of DeepSeek.