🚬 A new benchmark nof1.ai has appeared, where different LLMs trade cryptocurrency for real $10,000 to test intelligence in the markets.

In reality, the experiment looks absurd: models without news and context make random trades in the most volatile environment.

The results show nothing, leading is #Grok 4 simply because it doesn't trade — perhaps this is the only sign of true intelligence.

UPD: #DeepSeek has risen to the lead.