Qwen 3 Max Tops Nof1 AI Trading Test With 79% Gain

The assessment is 'likely_true' because the core numerical claim is supported by two separate sources, despite significant issues with source quality and conflicting data points.The only source that directly mentions the 'Nof1 AI Trading Test' and a 79% result (specifically, 79.43%) is a sponsored press release on coincu.com. This source has very low authority and high potential for bias, making it unreliable on its own.However, the claim is strongly corroborated by a more credible, independent AI comparison platform, vals.ai. This source shows Qwen 3 Max Preview achieving a 78.9% score on 'finance benchmarks'. The remarkable similarity between the 79.43% and 78.9% figures suggests they are very likely referring to the same underlying performance data, lending significant credibility to the numerical part of the statement.There is conflicting evidence from a LinkedIn post by vals-ai, which reports a low 17% accuracy for the model on a different, proprietary 'Finance Agent Benchmark'. This does not directly falsify the claim, as a model's performance can vary dramatically across different tests. It does, however, indicate that the 79% result is not representative of the model's performance on all financial tasks.Finally, the most authoritative sources, like the official Qwen GitHub repository and academic papers, make no mention of this test or result. This lack of official verification prevents a high confidence assessment.In summary, the statement is likely true in substance—that Qwen 3 Max achieved a score around 79% on a finance-related benchmark. The specific name 'Nof1 AI Trading Test' and the term 'gain' appear to be promotional framing from a low-quality source, but the underlying number is supported by more reliable data.