Grok 4 Delivered the Best Performance among 8 AI Models in Stock Market Trading

• Eight AI models were selected for the stock market trading project Rallies AI Arena.
• Among them, Grok 4 pulled ahead as the leader with a 7% profit.
• Meanwhile, Qwen 3, by contrast, lost more than $22,000.

---

The Tesla Owners Silicon Valley community reported that xAI’s Grok 4 chatbot is leading the interim results of a trading experiment. As of January 21, 2026, eight leading artificial intelligence (AI) models are competing against each other in the stock market.

As part of the Rallies AI Arena project, each model received $100,000 and has been trading stocks in real time since late November 2025, without any restrictions or “safety nets.”

The post notes:

The model’s current return is about +7%, which corresponds to a profit of $6,979 on the starting capital.

Tesla Owners SV emphasized that Grok 4 is not only ahead of its competitors, but is also outperforming broad market benchmarks under a real-money test.

Second and third place went to Anthropic’s Claude Sonnet 4.5 with a result of +5.7%, and Opus 4.5 with a return of +4.8%, respectively.

Next in the rankings were DeepSeek V3 (+3.3%) and Gemini 2.5 Pro (+2.9%).

GPT 5.2 and GPT 5.1 took sixth and seventh place, posting much more modest results, while Qwen 3 recorded a significant loss (more than $22,000). Entrepreneur and host of one of the biggest shows on the X platform, Mario Nawfal, commented on the situation as follows:

Recall that earlier, Grok 4.20 posted returns of 10–12% during Alpha Arena Season 1.5 — another live stock-trading competition. At the time, the model also operated in real market conditions and managed to outperform all competitors in terms of financial results.

At the same time, Grok’s competitor — Anthropic — has focused on the philosophical and ethical aspects of AI development. It published an updated version of the Constitution Claude document, which sets out the operating principles and limitations of the chatbot of the same name. Anthropic noted that the new revision places greater emphasis on safety, benefits for people, and the ethical boundaries of AI use.