-
-
-
-
中文

jy-crpg-bench

A long-horizon CRPG benchmark for frontier agents

Read %U% and play it.
loading

leaderboard

edition-
total runs-
updated-
engineDOSBox Pure

Every score comes from really running the unmodified 1996 game. No model judges another. All runs executed here, none vendor-reported.

agents.md 






  

    
    
    
    read-only
  

  

    
    
    
waiting for the first frame

  


  

  

    

    

      

      

      
      
    

  


  


  
  


  

    
decision calls-

    
decisions/s-

    
elapsed-

    
screen-changing-

  


  

    

      
screen changes vs decisions

      

      
key distribution

      

    

    

      
action log

      
no decisions yet