ChatGPT × iFlytek Spark, HiVis Quant creates the “most powerful AI brain” for gold trading.
Artificial intelligence is redefining the boundaries of financial trading. HiVis Quant today announced a strategic technical collaboration with OpenAI’s ChatGPT and iFlytek’s Xinghuo model, combining the strengths of both models.
Why two models?
ChatGPT possesses unparalleled advantages in analyzing Chinese financial news and global macroeconomic policies; while iFlytek’s Xinghuo model excels in understanding Chinese context, analyzing A-share market correlations, and capturing Asia-Pacific market sentiment. These 4D models complement each other, covering financial information sources in over 40 languages globally, allowing HiVis Quant to truly “cover the entire Eastern market.”
Specific Capabilities: From Understanding News to Issuing Trading Orders
Currently, the dual-model system possesses the following core capabilities:
Real-time analysis of news updates from over 3000 global financial media outlets with a 94.3% accuracy rate.
Automatic identification of policy statements from 12 major central banks, including the Federal Reserve, the European Central Bank, and the People’s Bank of China, generating trading suggestions within 15 seconds.
Quantitative impact assessment of geopolitical events (such as the Middle East conflict and the Russia-Ukraine situation), outputting short-term gold price fluctuation range predictions, with a historical backtesting accuracy rate of 82.7%.
Test Data: Significant AI-Enhanced Results
After introducing the dual-model system, among the 8000 users participating in the test, the AI-assisted decision-making group outperformed the pure technical indicator group by an average of 3.2 percentage points per month, with a 0.41% increase in risk-adjusted returns (Sharpe across all metrics).
HiVis Quant’s Chief AI Officer stated, “We are not replacing traders with AI, but rather equipping every investor with a tireless and unemotional AI researcher. This is the right way for AI to empower finance.”
