Wallet V don publish public benchmark for AI trading agents for Hyperliquid & Aster
Wallet V, one self-custody Web3 wallet, don launch public performance benchmark for AI trading agents wey users configure for decentralized derivatives platforms Hyperliquid and Aster. Di benchmark dey for Wallet V website and e dey publish aggregated cohort results.
Wallet V analyze 688 AI trading agents wey dem create over di past two months. Every agent na user-configured, e use model wey user choose (large language model) to generate trading decisions, then e execute trades for Hyperliquid or Aster. Wallet V dey aggregate performance by underlying model family, and dem go dey refresh results as new agents deploy.
Key stats: 42% of agents get profit/loss balance at or above zero. Peak agent-level ROI range from -30% (lowest-performing model) to +307% (highest-performing model). Model families wey get less than 10 agents dem treat as directional, no be statistically conclusive.
Across di cohort, agents trade perpetual futures across four asset buckets: major crypto assets (BTC, ETH, SOL), equity exposure (including pre-IPO), commodity benchmarks (gold, silver, oil), and major FX pairs. Instruments dem dey access via third-party venues.
Adam Cai (Founder & CEO of Virgo Group) talk say di benchmark give users institution-like way to evaluate models based on observable performance. Wallet V dey plan more releases, including more model families, prediction market support, advanced analytics for copilot trading, and personalized AI prompt generation.
For traders, di release add new public dataset to compare AI trading agents and model families, but di performance dispersion (from -30% to +307% ROI) show say results fit depend plenty on market regime.
Neutral
Di launch na na be mostly upgrade for data/benchmark, no be direct liquidity or token-issuance event. By to publish aggregated results from 688 AI trading agents and show say ROI disperse wide (from -30% to +307%), Wallet V fit make traders dey more confident for how dem go take choose models, but e still show say AI trading performance fit vary scatter by model family and market regime.
For short term, traders fit dey lean to monitor the benchmark when dem dey evaluate AI-driven strategies for Hyperliquid/Aster, wey fit increase attention and use of some model families. For long term, if the dataset prove say e dey predictive and refresh cadence remain steady, e fit slowly make AI-agent evaluation for onchain derivatives more legitimate.
Because the news no talk about new token incentives, protocol changes, or guaranteed performance, the net effect on broader market stability likely limited—so the stance na neutral not clearly bullish or bearish.