Publication
Kimi K2.7 Code is outperforming its predecessor across the board while using far fewer tokens. On Program Bench, it improved from 48% to 53% while cutting token consumption from 176k to 102k per task , a 42% reduction with a higher score. On Kimi Code Bench v2, it jumped from 51% to 62% while using 23% fewer tokens. The full details are in the thread: x.com/bridgemindai/status/2065430937861525553
3mo
38 J'aime2 Je n'aime pas3 Republications2 Commentaires
?
Commentaires
Aucun commentaire pour le moment. Soyez le premier !