Alpha Enviornment’s dwell crypto buying and selling benchmark confirmed Deepseek Chat V3.1 in first place on Saturday, October 18th, with the day’s leaderboard highlighting a slight rise on the prime and a drawdown for many rivals.
Deepseek tops leaderboard in Alpha Enviornment actual cash crypto battles
Deepseek Chat V3.1 led the pack with a Hyperliquid account worth of $10,400 (+4.0% return) after three accomplished trades. The bot paid a payment of $58.51, recorded a 0% win fee on closed trades, and recorded a most lack of $348.33 in opposition to a small unfavourable “win” of $4.19. This displays that energetic unrealized positions should not counted till they’re closed.
Grok-4 was in second place with $10,010 (+0.1%) and $0 charges, with no accomplished trades recorded as much as the snapshot. Claude Sonnet 4.5 ranked third with $9,985 (-0.15%) with $42.63 in charges and three closed trades with a most lack of $88.38. The experiment highlights how dramatically synthetic intelligence (AI) has superior lately.
Supply: nof1.ai leaderboard.
GPT-5 took 4th place with $9,901 (-0.99%) after two closes and $10.10 in charges, with a most lack of $59.04. Gemini 2.5 Professional ranked fifth with $9,725 (-2.75%) and paid the best fee of the day ($106.46) on 5 trades. Though this was the most important single win of the day ($329.35), it additionally represented a major lack of $731.43 and a 60% win fee on closed positions.
Qwen3 Max closed the sphere at $9,474 (-5.26%), together with $44.62 in charges and one closed commerce. The mannequin’s most win and loss are each -$517.77, indicating that there’s one notable shedding consequence. Total, Sharp’s readings had been low or unfavourable, in step with a restricted variety of trades and early spherical noise fairly than steady risk-adjusted efficiency.
Alpha Enviornment, launched by analysis institute Nof1 on October 17, allocates $10,000 to every mannequin to autonomously commerce crypto perpetual trades on the Hyperliquid decentralized trade (DEX). The Alpha Enviornment public dashboard tracks account worth, returns, whole P&L, commissions, win share, max win/loss, sharps, and trades whereas excluding unrealized P&L till a place is closed. This is a vital consideration when deciphering the every day standings.
Saturday’s snapshot on the Nof1.ai leaderboard reveals the experiment’s assumptions: similar price range, totally different LLM inference, and clear execution. Some bots are displaying zero or only a few accomplished trades, so the preliminary rank could fluctuate as open positions are resolved and the payment footprint will increase. For now, Deepseek holds the sting, however Grok-4’s clear slate retains it shut, and Gemini’s extraordinary win-loss mixture highlights the upper disparity.
FAQ
- What’s Alpha Enviornment? A dwell benchmark the place 6 LLMs autonomously commerce crypto perpetual trades for $10,000 every.
- Which mannequin led on October 18th? Deepseek Chat V3.1 led with $10,400 (+4.0%) based mostly on accomplished trades.
- The place do transactions happen? Hyperliquid decentralized trade with clear on-chain monitoring.
- Does the rating embody open P&L? No, solely accomplished transactions are counted. Lively positions replace their ranks when they’re closed.
