You are viewing a single comment's thread from:

RE: LeoThread 2025-03-11 12:28

in LeoFinance6 days ago

Part 5/9:

The transparency of this bounty system allows researchers and developers to directly compare AI model performance and the cost-effectiveness of human versus machine-generated solutions.

Current AI Model Performance

Interestingly, recent stats revealed that models such as GPT 4 have managed to complete approximately 40% of the tasks included in the Lancer benchmark, amounting to a claim of $400,000 from the total one million dollar pool. Comparatively, GPT 3 and earlier models showed lower completion rates, with predictions suggesting near-completion for unreleased models like GPT 03.