±¾ÔÂÏà¹Ø²¿·Öת´ïÖ÷ÒªÑо¿Ð§¹û³éѪ²éÖ×Áö±ê¼ÇÎï¿ÉÌæ»»³¦¾µÂð ºÜÐË·ÜΪÄú½â´ðÕâ¸öÎÊÌ⣬£¬£¬£¬£¬ÈÃÎÒÀ´°ïÄúÏêϸ˵Ã÷һϡ£¡£¡£¡£¡£Æ·ÅÆÊÚȨ±¨Ð޵绰£¬£¬£¬£¬£¬¿ìËÙÉÏÃÅ·þÎñ

ɽÎ÷Ê¡ÑôȪÊÐƽ¶¨ÏØÖØÇìÊÐÊÐϽÇø±±íÕÇø½­Î÷Ê¡¾°µÂÕòÊвý½­Çø¾®ÚêÏØÎâ¼ÒÒ¤ÏçÎ÷ÏçÌÁÇøÃÜÔÆÇøºÓÄÏÕ¯ÕòÎäÇåÇøãâ¹Á¸ÛÕòËÄ´¨Ê¡Ò˱öÊн­°²ÏØÁÉÄþÊ¡´óÁ¬ÊÐÂÃ˳¿ÚÇøºÍÌïµØÇøÑÓÇìÇø¿µ×¯ÕòÉÂÎ÷Ê¡±¦¼¦ÊÐǧÑôÏØÔÆÄÏÊ¡ºìºÓ¹þÄá×åÒÍ×å×ÔÖÎÖÝãòÎ÷ÏØ°ö²ºÊаöɽÇø½­Î÷ʡӥ̶ÊÐÓ཭ÇøÒø´¨ÊкØÀ¼ÏØÇå¾²ÇøÄÏÊнֵÀɽ¶«Ê¡ÁĶ¼»áÜÝƽÇø¾²º£ÇøÎ÷µÔׯÕòÐÐÌÆÏØÁúÖÝÕòÁÉÄþÊ¡ÉòÑôÊлëÄÏÇø¼»ÖÝÇø¶«¶þÓªÕòËÄ´¨Ê¡µÂÑôÊй㺺ÊÐÎ÷²ØɽÄÏÊм»ÖÝÇø¹ÙׯÕòºÚÁú½­Ê¡Æß̨ºÓÊÐÌÒɽÇøºÓ¶«Çø¶«Ð½ֵÀɽÎ÷Ê¡½ú¶¼»á¸ßƽÊаö²ºÊаöɽÇøÃÜÔÆÇø²»ÀÏÍÍÕòÔÆÄÏÊ¡ÎÄɽ׳×åÃç×å×ÔÖÎÖݹãÄÏÏغÍÌïµØÇø³ç×óÊи£½¨Ê¡ÆÎÌïÊгÇÏáÇøƽɽÏظÚÄÏÕò½­Î÷Ê¡¼ª°²ÊÐÓÀÐÂÏؼ»ÖÝÇø¶«Ê©¹ÅÕòºþ±±Ê¡¶÷Ê©ÍÁ¼Ò×åÃç×å×ÔÖÎÖݶ÷Ê©Êкþ±±Ê¡Ò˲ýÊÐÒÄÁêÇøËÄ´¨Ê¡Ò˱öÊдäÆÁÇøºÚÁú½­Ê¡º×¸ÚÊй㶫ʡÉÇÍ·ÊÐÄÏ°ÄÏغþ±±Ê¡ÏåÑôÊÐÏåÖÝÇøÖØÇìÊÐÊÐϽÇø±±íÕÇøɽÎ÷Ê¡½úÖÐÊÐÁéʯÏØÎ÷²Ø°¢ÀïµØÇøÈÕÍÁÏعÌÔ­ÊÐÎ÷¼ªÏØËÄ´¨Ê¡µÂÑôÊй㺺ÊиÊËàÊ¡ÇìÑôÊÐÕòÔ­Ïؽ­Î÷Ê¡¾°µÂÕòÊвý½­ÇøËÄ´¨Ê¡Á¹É½ÒÍ×å×ÔÖÎÖݸÊÂåÏØÃÜÔÆÇø¹Å±±¿ÚÕòºÚÁú½­Ê¡¹þ¶û±õÊаÍÑåÏظÊËàÊ¡°×ÒøÊо°Ì©Ïغþ±±Ê¡¶÷Ê©ÍÁ¼Ò×åÃç×å×ÔÖÎÖݶ÷Ê©Êзą́ÇøºÍÒå½ÖµÀÉÂÎ÷Ê¡±¦¼¦ÊÐǧÑôÏØÔÆÄÏÊ¡ÆÕ¶ýÊо°¶«ÒÍ×å×ÔÖÎÏØÔÞ»ÊÏØԺͷÕòºÚÁú½­Ê¡¼¦Î÷Êм¦¹ÚÇøºÚÁú½­Ê¡ÒÁ´ºÊнðÁÖÇøÔÞ»ÊÏØԺͷÕò°¢¿ËËÕµØÇø°Ý³ÇÏØËÄ´¨Ê¡Á¹É½ÒÍ×å×ÔÖÎÖݸÊÂåÏØÇຣʡº£ÄϲØ×å×ÔÖÎÖݹóµÂÏØɽ¶«Ê¡±õÖÝÊгç×óÊÐÄþÃ÷ÏزýƽÇøСÌÀɽÕò³¤°²Çø¹ã°²½ÖµÀÒÁÀç¹þÈø¿Ë×ÔÖÎÖÝÕÑËÕÏغÚÁú½­Ê¡º×¸ÚÊÐÄÚÃɹÅÎýÁÖ¹ùÀÕÃËÕýÏâ°×ÆìƽɽÏØÕ¬±±Ïç¹óÖÝÊ¡Ç­Äϲ¼ÒÀ×åÃç×å×ÔÖÎÖÝ»ÝË®ÏØÃÅÍ·¹µÇø´óÓø½ÖµÀɽÎ÷Ê¡ÂÀÁºÊÐá°ÏغӶ«Çø´óÖ±¹Á½ÖµÀºÓ¶«Çø´óÖ±¹Á½ÖµÀËÄ´¨Ê¡Ò˱öÊÐçîÏØ·À³Ç¸ÛÊж«ÐËÊÐÇÅÎ÷ÇøÁôÓª½ÖµÀ½¹×÷Êзą́ÇøÓÒ°²ÃŽֵÀɽÎ÷Ê¡ÁÙ·ÚÊа²ÔóÏØɽÎ÷Ê¡Ë·ÖÝÊÐɽÒõÏØÂåÑôÊÐÈêÑôÏزýƽÇøСÌÀɽÕò»³ÈáÇøÁúɽ½ÖµÀɽÎ÷Ê¡½úÖÐÊÐÁéʯÏØ´óÐËÇø±¦ÛæÇøÅ£¼ÒÅÆÕòËÄ´¨Ê¡³É¶¼»á½ðÅ£ÇøºÚÁú½­Ê¡¼Ñľ˹ÊÐÌÀÔ­ÏØÎ÷³ÇÇøÌìÇŽֵÀ±¦ÛæÇøÅ£¼ÒÅÆÕòÎ÷³ÇÇøÌìÇŽֵÀÁÉÄþÊ¡ÉòÑôÊÐÁÉÄþÊ¡ÌúÁëÊÐÇåºÓÇø¼ªÁÖÊ¡³¤´ºÊÐË«ÑôÇøÔÆÄÏÊ¡´óÀí°××å×ÔÖÎÖÝÔÆÁúÏØ

½ñÈÕ¹Ù·½ÇþµÀÅû¶ÐÐÒµ¶¯Ì¬±¾ÖÜ AI ÏîÄ¿ÍƼö£ºWorldModelBench£¬£¬£¬£¬£¬WorldScore£¬£¬£¬£¬£¬WorldLens¡­ÌìÏÂÄ£×Ó¾®Å磬£¬£¬£¬£¬ÐèҪеÄBenchmark£¬£¬£¬£¬£¬ºÜÐË·ÜΪÄú½â´ðÕâ¸öÎÊÌ⣬£¬£¬£¬£¬ÈÃÎÒÀ´°ïÄúÏêϸ˵Ã÷Ò»ÏÂ:¹Ù·½·þÎñרÏߣ¬£¬£¬£¬£¬Ö§³Ö¶àÆ·ÅƱ¨ÐÞ

ÃÜÔÆÇø¸ßÁëÕòº£ÄÏÊ¡ÈýɳÊÐÎ÷ɳÇøɽÎ÷Ê¡ÐÃÖÝÊÐÄþÎäÏØɽÎ÷Ê¡Ô˶¼»áÏÄÏØÇÅÎ÷Çø¶«»ª½ÖµÀº£ÄÏÊ¡ÈýɳÊÐÎ÷ɳÇøÒø´¨ÊкØÀ¼ÏØÐÅÑôÊÐƽÇÅÇøÉÂÎ÷Ê¡ººÖÐÊÐÄÏÖ£ÇøËÄ´¨Ê¡¸Ê×βØ×å×ÔÖÎÖݾÅÁúÏظÊËàÊ¡¶¨Î÷ÊÐμԴÏØÇຣʡ¹ûÂå²Ø×å×ÔÖÎÖÝ°ö²ºÊаöɽÇøº£ÄÏÊ¡º£¿£¿£¿£¿£¿£¿ÚÊÐÃÀÀ¼ÇøËÄ´¨Ê¡ËìÄþÊÐÅîϪÏظÊËàÊ¡°×ÒøÊо°Ì©Ïؽ­ËÕÊ¡Õò½­ÊоäÈÝÊÐÏòÑôÇø¶¹¸÷ׯµØÇøÇÅÎ÷Çø¶«»ª½ÖµÀÎÚ³ľÆëÊÐɳÒÀ°Í¿ËÇø¸ÊËàÊ¡ÌìË®ÊÐÇØÖÝÇø¹ã¶«Ê¡ÉÇÍ·ÊÐÁúºþÇøɽÎ÷Ê¡ÐÃÖÝÊÐÄþÎäÏغÓÎ÷ÇøÌÒÔ°½ÖµÀ¸ÊËàÊ¡ÌìË®ÊÐÇØÖÝÇø±±º£Êк£³ÇÇøÕÔÏØɳºÓµêÕòÏòÑôÇøС¹Ø½ÖµÀ½­ËÕÊ¡Á¬ÔƸÛÊиÓÓÜÇøËÄ´¨Ê¡ÀÖɽÊÐÁÉÄþÊ¡½õÖÝÊÐÒåÏؽ­ËÕÊ¡ÄϾ©ÊÐÆÖ¿ÚÇøÁÉÄþÊ¡ÁÉÑôÊа×ËþÇøÇຣʡÓñÊ÷²Ø×å×ÔÖÎÖÝÖζàÏØÇÅÎ÷ÇøÔ·¶«½ÖµÀɽ¶«Ê¡ÔæׯÊÐ̨¶ùׯÇøÃÜÔÆÇøºÓÄÏÕ¯ÕòɽÎ÷Ê¡Ô˶¼»áÏÄÏØ·ą́ÇøÓÒ°²ÃŽֵÀÇàÐãÇø¼»ÖÝÇø¹ÙׯÕò¹ðÁÖÊÐÐã·åÇøɽÎ÷Ê¡ÂÀÁºÊÐá°Ïؼ»ÖÝÇø¶«Ê©¹ÅÕòÏòÑôÇøС¹Ø½ÖµÀÔÆÄÏÊ¡ºìºÓ¹þÄá×åÒÍ×å×ÔÖÎÖÝãòÎ÷ÏØÑÓÇìÇø¿µ×¯ÕòÎ÷³ÇÇøÌìÇŽֵÀ¸ÊËàʡ¤ÄÏÊÐÎ䶼ÇøËÄ´¨Ê¡ÀÖɽÊйÌÔ­ÊÐÎ÷¼ªÏØÎ÷³ÇÇøÌìÇŽֵÀ»³ÈáÇøÑãÆܵØÇøº£µíÇøÇàÁúÇŽֵÀ¼»ÖÝÇø¶«¶þÓªÕòËÄ´¨Ê¡Á¹É½ÒÍ×å×ÔÖÎÖݸÊÂåÏØ°ö²ºÊаöɽÇø¼»ÖÝÇø¶«Ê©¹ÅÕò˳ÒåÇø¿Õ¸Û½ÖµÀ¸£½¨Ê¡ÄÏƽÊÐɽ¶«Ê¡µÂÖÝÊÐÆëºÓÏØÐÐÌÆÏØÁúÖÝÕò½¹×÷ÊÐÖÐÕ¾ÇøËÄ´¨Ê¡Ò˱öÊн­°²Ïؾ²º£Çø´óÇñׯÕò¿¦Ê²µØÇøÂó¸ÇÌáÏؽ­Î÷Ê¡¸ÓÖÝÊÐʯ³ÇÏعðÁÖÊÐÐã·åÇø½­ËÕÊ¡ÐìÖÝÊзáÏؽ­Î÷Ê¡¾Å½­ÊÐÎäÄþÏØËÄ´¨Ê¡ËìÄþÊÐÅîϪÏعȪÇø°×¹ȪÏç¹óÖÝÊ¡Ç­Äϲ¼ÒÀ×åÃç×å×ÔÖÎÖÝ»ÝË®ÏØ°ÙÉ«ÊÐÌïÁÖÏØɽÎ÷Ê¡½ú¶¼»áÔóÖÝÏع㶫ʡ»ÝÖÝÊÐÁúÃÅÏØÄÚÃɹŶõ¶û¶à˹Êпµ°ÍʲÇøÄÚÃɹÅÎýÁÖ¹ùÀÕÃËÕýÏâ°×Æì¼»ÖÝÇø¶«ÕÔ¸÷ׯÕòºÏ·ÊÊЮÑôÇøºþ±±Ê¡ÏåÑôÊÐÏåÖÝÇøºÚÁú½­Ê¡´óÐË°²ÁëµØÇøºôÂêÏØËÄ´¨Ê¡¸Ê×βØ×å×ÔÖÎÖݹóÖÝÊ¡Ç­Äϲ¼ÒÀ×åÃç×å×ÔÖÎÖÝ»ÝË®ÏØ¡°²ÏØÔÆÄÏÊ¡ºìºÓ¹þÄá×åÒÍ×å×ÔÖÎÖÝÂÌ´ºÏØÃÜÔÆÇø¹Å±±¿ÚÕòÎ÷²ØɽÄÏÊкÚÁú½­Ê¡Ë绯ÊÐÃ÷Ë®ÏØËÄ´¨Ê¡¹ã°²ÊÐÔÀ³ØÏس¤°²Çø¹ã°²½ÖµÀɽ¶«Ê¡ÑĮ̀ÊÐÁú¿ÚÊÐÉÂÎ÷Ê¡ººÖÐÊÐÁô°ÓÏØɽÎ÷Ê¡ÂÀÁºÊÐá°ÏØɽ¶«Ê¡Î«·»ÊÐÇàÖÝÊÐɽ¶«Ê¡×Ͳ©ÊÐÁÙ×ÍÇøÔÆÄÏÊ¡Àö½­ÊÐÄþÝõÒÍ×å×ÔÖÎÏعóÖÝÊ¡ÁùÅÌË®ÊÐË®³ÇÏØÁÉÄþÊ¡ÌúÁëÊпªÔ­ÊÐÇÅÎ÷Çø¶«»ª½ÖµÀ

È«Çò·þÎñÇøÓò:¸£½¨Ê¡ÈýÃ÷ÊÐ÷ÁÐÇø¼»ÖÝÇø¶«Ê©¹ÅÕò°¢¿ËËÕµØÇø°Ý³ÇÏغ£ÄÏÊ¡º£¿£¿£¿£¿£¿£¿ÚÊÐÃÀÀ¼Çø²ýƽÇøСÌÀɽÕòºþ±±Ê¡Ò˲ýÊÐïö¹éÏØ°¢¿ËËÕµØÇø°Ý³ÇÏØÉÂÎ÷Ê¡ººÖÐÊÐÁô°ÓÏØɽÎ÷Ê¡Ë·ÖÝÊÐɽÒõÏØ°¢¿ËËÕµØÇøкÍÏØÎ÷ÇàÇø¾«ÎäÕòÎ÷³ÇÇøÔÂ̳½ÖµÀÏòÑôÇø¹ÜׯµØÇø¸ÊËàʡ¤ÄÏÊÐÎ䶼ÇøƽɽÏض«»ØÉáÕòɽÎ÷Ê¡½úÖÐÊÐÌ«¹ÈÇø¹óÖÝÊ¡ÁùÅÌË®ÊÐË®³ÇÏØ°¢¿ËËÕµØÇøкÍÏظ£½¨Ê¡¸£ÖÝÊÐÓÀÌ©ÏØÔÞ»ÊÏØÎ÷ÁúÃÅÏçɽ¶«Ê¡ÇൺÊгÇÑôÇø½­Î÷Ê¡ÒË´ºÊÐÒË·áÏØÉÂÎ÷Ê¡ÏÌÑôÊÐÐËƽÊи£½¨Ê¡ÈýÃ÷ÊÐ÷ÁÐÇøËÄ´¨Ê¡Ò˱öÊдäÆÁÇø¹ȪÇø°×¹ȪÏç¹ÌÔ­ÊÐÎ÷¼ªÏØËÄ´¨Ê¡ÀÖɽÊи£½¨Ê¡¸£ÖÝÊÐÓÀÌ©ÏØÁÉÄþÊ¡ÏòÑôÊб±Æ±ÊиÊËàÊ¡ÇìÑôÊÐÕòÔ­ÏغӶ«Çø¶«Ð½ֵÀÏòÑôÇø¶¹¸÷ׯµØÇøÏòÑôÇøС¹Ø½ÖµÀ¾®ÚêÏØÎâ¼ÒÒ¤Ïç¼»ÖÝÇø¶«Ê©¹ÅÕòƽ¶¥É½ÊÐÎÀ¶«ÇøÃÅÍ·¹µÇø´óÓø½ÖµÀÎ÷²Ø°¢ÀïµØÇøÈÕÍÁÏØ°ÙÉ«Êо¸Î÷ÊÐÇÕÖÝÊк£ÄÏÊ¡ÈýɳÊÐÎ÷ɳÇøÔªÊÏÏØÒó´åÕòºþ±±Ê¡Ò˲ýÊÐÒ˶¼»á¹óÖÝÊ¡ÁùÅÌË®ÊÐË®³ÇÏعþÃÜÊÐÒÁÎáÏس¤°²ÇøÄÏ´åÕòÃÜÔÆÇø²»ÀÏÍÍÕòËÄ´¨Ê¡¸Ê×βØ×å×ÔÖÎÖÝãò¶¨ÏØÃÅÍ·¹µÇø´óÓø½ÖµÀËÄ´¨Ê¡ÀÖɽÊкþ±±Ê¡Ò˲ýÊÐïö¹éÏØÎäÇåÇøãâ¹Á¸ÛÕòºþ±±Ê¡ÏåÑôÊÐÏåÖÝÇøÎ÷ÇàÇø¾«ÎäÕòɽÎ÷Ê¡½úÖÐÊÐÌ«¹ÈÇø¹óÖÝÊ¡Í­ÈÊÊÐÓñÆÁ¶±×å×ÔÖÎÏØÇຣʡº£ÄϲØ×å×ÔÖÎÖݹóµÂÏØÇÅÎ÷Çø¶«»ª½ÖµÀɽ¶«Ê¡Ì©°²Êж«Æ½ÏØÁÉÄþÊ¡ÌúÁëÊÐÇåºÓÇøÉϺ£ÊÐÊÐϽÇø¼Î¶¨Çø²©¶ûËþÀ­ÃɹÅ×ÔÖÎÖÝÎÂȪÏØ˳ÒåÇø¿Õ¸Û½ÖµÀÇå¾²ÇøÄÏÊнֵÀ¾²º£ÇøÎ÷µÔׯÕòå§ÑôÊÐÄÏÀÖÏعðÁÖÊÐÐË°²Ïظ£½¨Ê¡ÆÎÌïÊгÇÏáÇø½­ËÕÊ¡ÄÏͨÊÐÆô¶«ÊÐÁÉÄþÊ¡½õÖÝÊÐÒåÏØɽÎ÷Ê¡½ú¶¼»á¸ßƽÊÐÔÆÄÏÊ¡ºìºÓ¹þÄá×åÒÍ×å×ÔÖÎÖÝãòÎ÷Ïؾ²º£Çø´óÇñׯÕò½­Î÷Ê¡¼ª°²ÊÐÓÀÐÂÏؽ­ËÕÊ¡ÄÏͨÊÐÆô¶«ÊиÊËàʡ¤ÄÏÊÐɽÎ÷Ê¡³¤ÖÎÊÐÏåÔ«ÏØ°ÙÉ«Êо¸Î÷ÊÐÏòÑôÇø¾ÆÏÉÇŽֵÀºþ±±Ê¡Ò˲ýÊÐÒ˶¼»áƽɽÏØС¾õÕòÃÜÔÆÇø²»ÀÏÍÍÕòËÄ´¨Ê¡µÂÑôÊй㺺ÊÐÉÂÎ÷Ê¡ººÖÐÊÐÎ÷ÏçÏظÊËàÊ¡ÇìÑôÊÐÕòÔ­ÏغÚÁú½­Ê¡¼Ñľ˹Êи»½õÊн­Î÷Ê¡ÒË´ºÊзîÐÂÏØËÄ´¨Ê¡³É¶¼»áж¼ÇøÔÞ»ÊÏØԺͷÕò¸£½¨Ê¡¸£ÖÝÊÐÂÞÔ´ÏØɽ¶«Ê¡ÑĮ̀ÊÐÁú¿ÚÊÐɽ¶«Ê¡ÁĶ¼»áÜÝƽÇøº£ÄÏÊ¡ÈýɳÊÐÎ÷ɳÇø¹ã¶«Ê¡ÉÇÍ·ÊÐÁúºþÇø³ç×óÊо²º£Çø´óÇñׯÕòÎ÷ÇàÇø¾«ÎäÕòÔÆÄÏÊ¡ÓñϪÊÐÐÂƽÒÍ×å´ö×å×ÔÖÎÏغ£ÄÏÊ¡º£¿£¿£¿£¿£¿£¿ÚÊÐÃÀÀ¼Çø

±¾ÖÜÐÐÒµ±¨¸æ¸üÐÂ×îÐÂÕþ²ß±¾ÖÜ AI ÏîÄ¿ÍƼö£ºWorldModelBench£¬£¬£¬£¬£¬WorldScore£¬£¬£¬£¬£¬WorldLens¡­ÌìÏÂÄ£×Ó¾®Å磬£¬£¬£¬£¬ÐèҪеÄBenchmark£¬£¬£¬£¬£¬ºÜÐË·ÜΪÄú½â´ðÕâ¸öÎÊÌ⣬£¬£¬£¬£¬ÈÃÎÒÀ´°ïÄúÏêϸ˵Ã÷Ò»ÏÂ:ÊÛºó·þÎñάÐÞÖÐÐĵ绰£¬£¬£¬£¬£¬Ö§³Ö¶àÇþµÀ·þÎñ

ÌìÏ·þÎñÇøÓò£º½­Î÷Ê¡¼ª°²Êмª°²Ïغþ±±Ê¡¶÷Ê©ÍÁ¼Ò×åÃç×å×ÔÖÎÖݶ÷Ê©Êк£ÄÏÊ¡º£¿£¿£¿£¿£¿£¿ÚÊÐÃÀÀ¼Çø¹ã¶«Ê¡ÉîÛÚÊÐËÄ´¨Ê¡³É¶¼»áж¼ÇøÉÂÎ÷Ê¡±¦¼¦ÊÐǧÑôÏØÉÂÎ÷Ê¡±¦¼¦ÊÐüÏØÎ÷ÏçÌÁÇøÙñÖÝÊÐÀûÐÁÏزýƽÇøÑô·»ÕòÄÚÃɹÅÎýÁÖ¹ùÀÕÃËÏâ»ÆÆìËÄ´¨Ê¡³É¶¼»á½ðÅ£Çø²ýƽÇøСÌÀɽÕò¼»ÖÝÇø¶«Ê©¹ÅÕòÐí²ýÊн¨°²ÇøÉÂÎ÷Ê¡ÉÌÂåÊÐɽÑôÏØÃÜÔÆÇø²»ÀÏÍÍÕò¼ªÁÖÊ¡ËÄƽÊÐÌúÎ÷ÇøÉÂÎ÷Ê¡ººÖÐÊÐÎ÷ÏçÏغÚÁú½­Ê¡º×¸ÚÊб±³½Çø¹ãÔ´½ÖµÀɽÎ÷Ê¡ÂÀÁºÊÐá°ÏØÉÂÎ÷Ê¡ÉÌÂåÊÐɽÑôÏØɽÎ÷Ê¡ÐÃÖÝÊж¨ÏåÏØËÄ´¨Ê¡Á¹É½ÒÍ×å×ÔÖÎÖÝÕѾõÏØɽÎ÷Ê¡½úÖÐÊÐÓÜÉçÏع㶫ʡÉîÛÚÊÐËÄ´¨Ê¡ÀÖɽÊи£½¨Ê¡ÄÏƽÊп¦Ê²µØÇøÂó¸ÇÌáÏظÊËàʡ¤ÄÏÊÐÎ䶼ÇøºÚÁú½­Ê¡º×¸ÚÊÐÇÕÖÝÊйðÁÖÊÐÐã·åÇø¾®ÚêÏØÎâ¼ÒÒ¤Ïçºþ±±Ê¡Ò˲ýÊÐïö¹éÏر¦ÛæÇøÅ£¼ÒÅÆÕò°ÙÉ«Êо¸Î÷ÊÐËÄ´¨Ê¡Ò˱öÊдäÆÁÇø²ýƽÇøСÌÀɽÕò½­Î÷Ê¡¼ª°²ÊÐÓÀÐÂÏØ°ö²ºÊÐÁú×ÓºþÇø¾®ÚêÏØÎâ¼ÒÒ¤Ïçɽ¶«Ê¡ÑĮ̀ÊÐÁú¿ÚÊÐɽ¶«Ê¡Ì©°²Êж«Æ½ÏØɽ¶«Ê¡×Ͳ©ÊÐÁÙ×ÍÇø³¤°²ÇøÄÏ´åÕòºþ±±Ê¡Ò˲ýÊÐÒ˶¼»á»³ÈáÇø±¦É½Õò½­Î÷Ê¡¾Å½­ÊÐÎäÄþÏØ»´ÄÏÊдóͨÇøɽ¶«Ê¡±õÖÝÊÐå§ÑôÊÐÄÏÀÖÏع㶫ʡ»ÝÖÝÊÐÁúÃÅÏؽ­ËÕÊ¡ÐìÖÝÊзáÏس¤°²Çø¹ã°²½ÖµÀºÚÁú½­Ê¡Ë绯ÊÐÃ÷Ë®ÏØÔÆÄÏÊ¡ºìºÓ¹þÄá×åÒÍ×å×ÔÖÎÖÝãòÎ÷ÏØ¡°²Ïع㶫ʡÉÇÍ·ÊÐÁúºþÇø¹óÖÝÊ¡Ç­Äϲ¼ÒÀ×åÃç×å×ÔÖÎÖÝ»ÝË®ÏØ»³ÈáÇøÑãÆܵØÇøɽ¶«Ê¡µÂÖÝÊÐÆëºÓÏØÒÁÀç¹þÈø¿Ë×ÔÖÎÖÝÕÑËÕÏØÔÆÄÏÊ¡ÓñϪÊÐÐÂƽÒÍ×å´ö×å×ÔÖÎÏØÉÂÎ÷Ê¡ººÖÐÊÐÎ÷ÏçÏعðÁÖÊÐÐË°²ÏØÉÂÎ÷Ê¡ÏÌÑôÊÐÐËƽÊкÚÁú½­Ê¡Ë绯ÊÐÃ÷Ë®Ïؼ»ÖÝÇø¶«ÕÔ¸÷ׯÕòËÄ´¨Ê¡Á¹É½ÒÍ×å×ÔÖÎÖݸÊÂåÏØÉÂÎ÷Ê¡ººÖÐÊÐÁô°ÓÏظ£½¨Ê¡ÆÎÌïÊÐÏÉÓÎÏØÎäÇåÇøÉÏÂį́Õòº£ÄÏÊ¡º£¿£¿£¿£¿£¿£¿ÚÊÐÃÀÀ¼Çø³ç×óÊÐÇຣʡº£ÄϲØ×å×ÔÖÎÖݹóµÂÏØÄÏÑôÊÐÄÚÏçÏغþ±±Ê¡ÏåÑôÊÐÏåÖÝÇøÎ÷³ÇÇøÔÂ̳½ÖµÀºþ±±Ê¡Ð¢¸ÐÊкº´¨ÊÐÐÂÀÖÊÐЭÉñÏç½­ËÕÊ¡ËÕÖÝÊÐÏà³ÇÇø¹ã¶«Ê¡ÉîÛÚÊÐÁÉÄþÊ¡´óÁ¬ÊÐÂÃ˳¿ÚÇøËÄ´¨Ê¡³É¶¼»á½ðÅ£ÇøËÄ´¨Ê¡ËìÄþÊÐÅîϪÏØÔÆÄÏÊ¡´óÀí°××å×ÔÖÎÖÝÔÆÁúÏØÏòÑôÇøË«¾®½ÖµÀÃÜÔÆÇø¸ßÁëÕò½­ËÕÊ¡ËÕÖÝÊÐÏà³ÇÇøÇÅÎ÷Çø¶«»ª½ÖµÀ¹ã¶«Ê¡ÉÇÍ·ÊÐÄÏ°ÄÏر±º£Êк£³ÇÇø¹ã¶«Ê¡ÔƸ¡ÊÐÐÂÐËÏØÎ÷ÏçÌÁÇøº£ÄÏÊ¡ÙÙÖÝÊгç×óÊÐɽÎ÷Ê¡Ë·ÖÝÊÐɽÒõÏØËÄ´¨Ê¡¹ãÔªÊÐ

ÊÛºó·þÎñÉÏÃÅ·þÎñµç»°£¬£¬£¬£¬£¬ÖÇÄÜ·ÖÅäƱ¾Ý£º±¾ÖÜ AI ÏîÄ¿ÍƼö£ºWorldModelBench£¬£¬£¬£¬£¬WorldScore£¬£¬£¬£¬£¬WorldLens¡­ÌìÏÂÄ£×Ó¾®Å磬£¬£¬£¬£¬ÐèҪеÄBenchmark

×÷Õߣü·®ÑÅæÃ΢ÐÅ£ü FFF111f__

2026 ÄêÉÏ°ëÄ꣬£¬£¬£¬£¬¡°ÌìÏÂÄ£×Ó¡±ÒѾ­´ÓÂÛÎÄÀïµÄÊÖÒÕ´Ê£¬£¬£¬£¬£¬Äð³ÉÁË AI ¹«Ë¾¡¢»úеÈ˹«Ë¾¡¢×Ô¶¯¼ÝÊ»¹«Ë¾ºÍ VC ÌÖÂÛÀïÈƲ»¿ªµÄÒªº¦´Ê¡£¡£¡£¡£¡£ÊÓƵÌìÉúÄ£×Ó˵×Ô¼ºÄÜÄ£ÄâÎïÀíÌìÏ£¬£¬£¬£¬£¬¾ßÉíÖÇÄÜÍŶÓ˵×Ô¼ºÒªÓÃÌìÏÂÄ£×ÓѵÁ·»úеÈË£¬£¬£¬£¬£¬×Ô¶¯¼ÝÊ»¹«Ë¾ËµÌìÉúʽ·ÂÕæ¿ÉÒÔÌæ»»Ò»²¿·ÖÕæʵ·²â¡£¡£¡£¡£¡£

µ«ÎÊÌâÒ²ËæÖ®±äµÃ¼âÈñ£ºÒ»¸öÄ£×Ó»áÌìÉúƯÁÁÊÓƵ£¬£¬£¬£¬£¬¾Í¼´ÊÇËüÃ÷È·ÌìÏÂÁËÂ𣿣¿£¿£¿£¿£¿ËüÄܲ»¿É×ñÊØÎïÀí¼ÍÂÉ£¿£¿£¿£¿£¿£¿Äܲ»¿É¼á³Ö 3D ¼¸ºÎÒ»Ö£¿£¿£¿£¿£¿£¿Äܲ»¿É·þÎñ×Ô¶¯¼ÝÊ»ºÍ»úеÈ˾öÒ飿£¿£¿£¿£¿£¿Äܲ»¿É´Ó¡°¿´ÆðÀ´ÏñÕæµÄ¡±£¬£¬£¬£¬£¬×ßµ½¡°ÓÃÆðÀ´ÕæµÄÓÐÓá±£¿£¿£¿£¿£¿£¿

ÒÔÊÇ£¬£¬£¬£¬£¬±ÈÆðÖ»¿´ demo£¬£¬£¬£¬£¬¸üÓ¦¸Ã¿´ benchmark¡£¡£¡£¡£¡£ÏÂÃæÕâÎå¸öÌìÏÂÄ£×Óbenchmark°üÀ¨Í¨ÓÃÊÓƵÌìÏÂÄ£×Ó¡¢Í³Ò»ÌìÏÂÌìÉú¡¢×Ô¶¯¼ÝÊ»¡¢¾ßÉíÖÇÄÜ¡¢ÎïÀíÍÆÀí¡£¡£¡£¡£¡£ËüÃDz»ÊÇËùÓÐ benchmark µÄÈ«¼¯£¬£¬£¬£¬£¬µ«»ù±¾ÁýÕÖÁ˵±ÏÂÌìÏÂÄ£×Ó×îÒªº¦µÄÎåÌõÖ÷Ïß¡£¡£¡£¡£¡£

1. WorldModelBench

Ò»¾ä»°ÏÈÈÝ£º WorldModelBench °ÑÊÓƵÌìÉúÄ£×Ó¿´³ÉÌìÏÂÄ£×ÓÀ´¿¼£¬£¬£¬£¬£¬²»µ«ÎÊ»­ÃæºÃÇ·ÔÃÄ¿£¬£¬£¬£¬£¬»¹ÎÊÖ¸Á֪ʶºÍÎïÀíÊÇ·ñ¾­µÃÆð¼ì²é¡£¡£¡£¡£¡£

WorldModelBench ½â¾öµÄÊÇͨÓÃÊÓƵģ×ÓÆÀ²âÀï×î½¹µãµÄȱ¿Ú£ºÒÑÍùÐí¶à benchmark Æ«ÏòÊÓ¾õÖÊÁ¿¡¢Îı¾¶ÔÆëijÈËÀàÆ«ºÃ£¬£¬£¬£¬£¬µ«¡°ÌìÏÂÄ£×Ó¡±ÕæÕýÒª»Ø¸²µÄÊÇ£¬£¬£¬£¬£¬Ä£×ÓÊÇ·ñÃ÷È·Ðж¯¡¢³¡¾°Òò¹ûºÍ»ù´¡ÎïÀí¡£¡£¡£¡£¡£ËüÁýÕÖ»úеÈË¡¢×Ô¶¯¼ÝÊ»¡¢¹¤Òµ¡¢ÈËÀà»î¶¯¡¢ÓÎÏ·¡¢¶¯»­¡¢×ÔÈ»µÈ 7 ¸öÓ¦ÓÃÓò£¬£¬£¬£¬£¬Ï¸·Öµ½ 56 ¸ö×ÓÓò£¬£¬£¬£¬£¬Ê¹Óà 350 ¸öͼÎÄÌáÐÑ£¬£¬£¬£¬£¬²¢ÓÃÖ¸Áî×·Ë桢֪ʶ¡¢ÎïÀí×ñÊØÈý´óά¶ÈÆÀ²â¡£¡£¡£¡£¡£ÎïÀí²¿·ÖÖص㿴ţ¶ÙµÚÒ»¶¨ÂÉ¡¢Ðα䡢Á÷Ìå¡¢´©Í¸¡¢ÖØÁ¦µÈ³£¼û¡°ÌìϽ¨Ä£»£»£»Ã¾õ¡±¡£¡£¡£¡£¡£ÂÛÎÄ»¹ÖÚ°üÁË 6.7 Íò¸öÈËÀà±ê×¢£¬£¬£¬£¬£¬²¢ÑµÁ·ÁËÒ»¸ö human-aligned judge À´×Ô¶¯ÆÀ²â¡£¡£¡£¡£¡£ÔÚЧ¹ûÉÏ£¬£¬£¬£¬£¬Kling µÈ±ÕÔ´Ä£×ÓÕûÌå¿¿Ç°£¬£¬£¬£¬£¬Mochi¡¢OpenSoraPlan µÈ¿ªÔ´Ä£×ÓÔÚ²¿·Öά¶È¿¿½ü±ÕÔ´Ä£×Ó£¬£¬£¬£¬£¬µ«ÂÛÎĵĽ¹µã½áÂÛ²»ÊÇ¡°Ë­ÒѾ­½â¾öÁËÌìÏÂÄ£×Ó¡±£¬£¬£¬£¬£¬¶øÊÇÍ·²¿Ä£×ÓÈÔÈ»Àë¿É¿¿ÌìÏÂÄ£ÄâÆ÷ºÜÔ¶¡£¡£¡£¡£¡£

Ðû²¼ÍŶÓÀ´×Ô UC Berkeley¡¢UC San Diego¡¢NVIDIA ºÍ MIT £¬£¬£¬£¬£¬Ion Stoica ÊÇ UC Berkeley ÅÌËã»ú½ÌÊÚ£¬£¬£¬£¬£¬Ò²ÊÇ Databricks¡¢Anyscale µÈ¹«Ë¾µÄÅäºÏÊ×´´ÈË£»£»£»Joseph E. Gonzalez ÊÇ Berkeley EECS ½ÌÊÚ¡¢Sky Computing Lab / RISE Lab ÅäºÏÈÏÕæÈË£¬£¬£¬£¬£¬²¢ÊôÓÚ BAIR Éú̬¡£¡£¡£¡£¡£Õâʹ WorldModelBench ²»µ«ÊÇÒ»¸öÊÓƵÆÀ·Ö¼¯£¬£¬£¬£¬£¬¶øÊÇϵͳ¡¢ÊÓ¾õ¡¢»úеÈ˺͸ßЧ AI ÍŶÓÅäºÏÍƶ¯µÄÌìÏÂÄ£×ÓÆÀ²â¡£¡£¡£¡£¡£

ΪʲôֵµÃ¹Ø×¢

ËüÊÇ×îÊÊÏàÖúΪ¡°Í¨ÓÃÊÓƵÌìÏÂÄ£×ÓÈëÃÅ»ù×¼¡±µÄ benchmark¡£¡£¡£¡£¡£ÓÉÓÚËüûÓÐֻͣÁôÔÚÉóÃÀºÍÇåÎú¶È£¬£¬£¬£¬£¬¶øÊÇ°ÑÌìÏÂÄ£×Ó²ð³ÉÓ¦ÓÃÓò¡¢ÎïÀí¼ÍÂɺÍϸÁ£¶È¹ýʧÀàÐÍ£¬£¬£¬£¬£¬ÌØÊâÊʺÏÅжÏÒ»¸öÊÓƵģ×Ó¾¿¾¹ÊÇÔÚ¡°ÌìÉúÏñÌìϵÄͼÏñ¡±£¬£¬£¬£¬£¬ÕÕ¾ÉÔÚ¡°Ä£ÄâÒ»¸öÄÜ×ÔÇ¢ÑÝ»¯µÄÌìÏ¡±¡£¡£¡£¡£¡£

ÏîÄ¿£ºhttps://worldmodelbench-team.github.io/

ÂÛÎÄ£ºhttps://arxiv.org/abs/2502.20694

´úÂëÓëÆÀ²â£ºhttps://github.com/WorldModelBench-Team/WorldModelBench

²âÊÔÊý¾Ý£ºhttps://huggingface.co/WorldModelBench-Team

2. WorldScore

Ò»¾ä»°ÏÈÈÝ£º WorldScore °Ñ 3D¡¢4D¡¢Í¼ÉúÊÓƵ¡¢ÎÄÉúÊÓƵģ×ӷŵ½Í³Ò»¸ö¡°ÏÂÒ»³¡¾°ÌìÉú¡±¿ò¼ÜÀïÆÀ²â£¬£¬£¬£¬£¬Öص㿴ģ×ÓÄܲ»¿É°´Ïà»ú¹ì¼£Ò»Á¬ÌìÉúÌìÏ¡£¡£¡£¡£¡£

¸ÅÊö£º WorldScore ½â¾öµÄÊÇ¡°²î±ðÌìÏÂÌìÉúõ辶ÎÞ·¨¹«Õý½ÏÁ¿¡±µÄÎÊÌâ¡£¡£¡£¡£¡£3D Ä£×Ó¡¢4D Ä£×ÓºÍÊÓƵģ×ÓÊä³öÐÎ̬²î±ð£¬£¬£¬£¬£¬ÒÑÍùºÜÄÑÔÚͳһÕÅÄÚÍâ½ÏÁ¿¡£¡£¡£¡£¡£WorldScore ½«ÌìÏÂÌìÉú²ð³ÉһϵÁлùÓÚÃ÷È·Ïà»ú¹ì¼£µÄ next-scene generation ʹÃü£¬£¬£¬£¬£¬Óà 3000 ¸öÑùÀýÁýÕÖ¾²Ì¬/¶¯Ì¬¡¢ÊÒÄÚ/ÊÒÍ⡢дʵ/Æø¸Å»¯µÈ³¡¾°£¬£¬£¬£¬£¬²¢´Ó¿É¿ØÐÔ¡¢ÖÊÁ¿¡¢¶¯Ì¬Èý·½Ãæ´ò·Ö¡£¡£¡£¡£¡£¿£¿£¿£¿£¿£¿É¿ØÐÔ¿´Ïà»ú¿ØÖÆ¡¢ÎïÌå¿ØÖƺÍÄÚÈݶÔÆ룬£¬£¬£¬£¬ÖÊÁ¿¿´ 3D Ò»ÖÂÐÔ¡¢¹â¶ÈÒ»ÖÂÐÔ¡¢Æø¸ÅÒ»ÖÂÐÔºÍÖ÷¹ÛÖÊÁ¿£¬£¬£¬£¬£¬¶¯Ì¬¿´Ô˶¯×¼È·ÐÔ¡¢·ù¶ÈºÍÁ÷ͨ¶È¡£¡£¡£¡£¡£¹Ù·½Ð§¹ûÏÔʾ£¬£¬£¬£¬£¬Voyager¡¢WonderWorld¡¢LucidDreamer µÈ 3D/³¡¾°ÌìÉúÄ£×ÓÔÚ¾²Ì¬ÌìÏÂÌìÉúÉϺÜÇ¿£¬£¬£¬£¬£¬CogVideoX-I2V µÈÊÓƵģ×ÓÔÚ¶¯Ì¬·ÖÊýÉÏÌåÏÖÍ»³ö£»£»£»Í¬Ê±£¬£¬£¬£¬£¬T2V Ä£×Óͨ³£¸üÈÝÒ×±»Îı¾¿ØÖÆ£¬£¬£¬£¬£¬I2V Ä£×ÓÍùÍù»­ÖʸüÎÈ¡£¡£¡£¡£¡£

WorldScore ÊÇ Stanford University ÍŶÓÐû²¼£¬£¬£¬£¬£¬ÂÛÎÄ×÷ÕßΪ Haoyi Duan¡¢Hong-Xing Yu¡¢Sirui Chen¡¢Li Fei-Fei¡¢Jiajun Wu¡£¡£¡£¡£¡£Li Fei-Fei ÊÇ Stanford ÅÌËã»úϵ Sequoia Professor¡¢Stanford HAI Ê×´´ÁªºÏÖ÷ÈΣ¬£¬£¬£¬£¬ÔøÈÎ Stanford AI Lab Ö÷ÈÎºÍ Google Cloud AI/ML Ê×ϯ¿Æѧ¼Ò£¬£¬£¬£¬£¬¿ª°ìWorld Labs£¬£¬£¬£¬£¬Ò²ÊÇ ImageNet µÄÒªº¦Íƶ¯Õߣ»£»£»Jiajun Wu ÊÇ Stanford ÅÌËã»úϵÖúÀí½ÌÊÚ£¬£¬£¬£¬£¬Í¬Ê±¼æÈÎÐÄÀíѧƫÏò£¬£¬£¬£¬£¬Ñо¿ÅÌËã»úÊÓ¾õ¡¢»úеÈ˺ÍÅÌËãÈÏÖª¿Æѧ£¬£¬£¬£¬£¬ËùÔÚÆ«Ïò×ÔÈ»Ìù½ü¡°ÌìÏÂÔõÑù±»ÌåÏÖ¡¢ÍÆÀíºÍÌìÉú¡±¡£¡£¡£¡£¡£Òò´Ë£¬£¬£¬£¬£¬WorldScore ¸üÏñÊÇ Stanford ÊÓ¾õѧϰ¡¢3D/ÎïÀí³¡¾°Ã÷È·ºÍÌìÉúÄ£×ÓÑо¿ÂöÂçϵÄÒ»Ì×ÌìÏÂÌìÉú»³±§¡£¡£¡£¡£¡£

ΪʲôֵµÃ¹Ø×¢£º

ÈôÊÇ˵ WorldModelBench ¸üÏñ¡°ÊÓƵģ×ÓÊDz»ÊÇÌìÏÂÄ£×Ó¡±µÄÌå¼ì£¬£¬£¬£¬£¬WorldScore Ôò¸üÏñ¡°ÌìÏÂÌìÉúõ辶´óºáÆÀ¡±¡£¡£¡£¡£¡£Ëü°Ñ 3D¡¢4D¡¢ÊÓƵÈýÌõÊÖÒÕõ辶·Å½øͳһЭÒ飬£¬£¬£¬£¬¶Ô´´Òµ¹«Ë¾ºÍÑо¿ÍŶÓÅжÏÊÖÒÕõ辶ºÜÓмÛÖµ£º¾¿¾¹¸ÃѺÊÓƵÌìÉú¡¢3D ÌìÉú£¬£¬£¬£¬£¬Õվɸü»ìÏýµÄ 4D ÌìÏÂÌìÉú¡£¡£¡£¡£¡£

ÏîÄ¿Ò³£ºhttps://haoyi-duan.github.io/WorldScore/

ÂÛÎÄ£ºhttps://arxiv.org/abs/2504.00983

´úÂ룺https://github.com/haoyi-duan/WorldScore

Êý¾Ý¼¯£ºhttps://huggingface.co/datasets/haoyi-duan/WorldScore

3. WorldLens

Ò»¾ä»°ÏÈÈÝ£º WorldLens רÃÅÆÀ¹À×Ô¶¯¼ÝÊ»ÌìÏÂÄ£×Ó£¬£¬£¬£¬£¬½¹µãÎÊÌâÊÇ£ºÌìÉúµÄ¼ÝÊ»Ììϲ»µ«ÒªÏñ£¬£¬£¬£¬£¬»¹Òª¼¸ºÎÎȹ̡¢ÎïÀíºÏÀí¡¢ÐÐΪÇå¾²¡¢¶ÔÏÂÓÎʹÃüÓÐÓᣡ£¡£¡£¡£

WorldLens ½â¾öµÄÊÇ×Ô¶¯¼ÝÊ»³¡¾°Àï¡°ÊÓ¾õ±ÆÕ浫¹¦Ð§Ê§Õ桱µÄÎÊÌâ¡£¡£¡£¡£¡£¼ÝÊ»ÌìÏÂÄ£×ÓºÜÈÝÒ×ÌìÉú¿´ÆðÀ´ºÏÀíµÄ 4D ³¡¾°£¬£¬£¬£¬£¬µ«Ò»µ©Çе½ÖØÐÞ¡¢±Õ»·¿ØÖÆ¡¢ÏÂÓμì²âÖ§½â£¬£¬£¬£¬£¬¾Í»á̻¶¼¸ºÎƯÒÆ¡¢Ô˶¯²»Îȹ̡¢Åöײ¡¢Æ«Àëõ辶µÈÎÊÌâ¡£¡£¡£¡£¡£WorldLens ÓÃÎå¸ö»¥²¹Æ«ÏòÆÀ²â£ºGeneration¡¢Reconstruction¡¢Action-Following¡¢Downstream Task¡¢Human Preference£¬£¬£¬£¬£¬²¢½øÒ»²½¹¹½¨ WorldLens-26K ÈËÀàÆ«ºÃÊý¾ÝºÍ WorldLens-Agent ÊÓ¾õÓïÑÔÆÀÉóÆ÷¡£¡£¡£¡£¡£ÏîÄ¿½áÂÛºÜÊÇÏÖʵ£ºÃ»Óмòµ¥Ä£×ÓÄÜÔÚËùÓÐά¶ÈͳÖΰñµ¥£¬£¬£¬£¬£¬ÎÆÀíÕæʵµÄÄ£×Ó¿ÉÄÜÎïÀíºÍ¼¸ºÎ²»ÎÈ£¬£¬£¬£¬£¬¼¸ºÎÎȹ̵ÄÄ£×ÓÓÖ¿ÉÄÜÐÐΪϸ½Úȱ·¦¡£¡£¡£¡£¡£ÂÛÎÄʵÑéÖУ¬£¬£¬£¬£¬DiST-4D ÔÚÐÂÊÓ½ÇÖÊÁ¿ºÍ²¿·ÖÏÂÓÎʹÃüÉÏÌåÏÖÇ¿£¬£¬£¬£¬£¬OpenDWM µÈÄ£×ÓÔÚ¼¸ºÎÎȹÌÐÔÉÏÓÐÓÅÊÆ£¬£¬£¬£¬£¬DriveDreamer-2 Ò²ÔÚʹÃüÏà¹ØÖ¸±êÖнøÈëÇ¿ÊÆÌݶÓ¡£¡£¡£¡£¡£

WorldLens ÓÉ WorldBench Team Ðû²¼£¬£¬£¬£¬£¬NUS¡¢University of Macau¡¢USTC¡¢Zhejiang University¡¢NTU¡¢Horizon Robotics¡¢HUST¡¢TUM¡¢Fudan University¡¢Shanghai AI Lab¡¢A*STAR¡¢CNRS µÈ¡£¡£¡£¡£¡£½¹µã×÷Õß°üÀ¨ Ao Liang¡¢Lingdong Kong¡¢Tianyi Yan¡¢Hongsi Liu¡¢Ziqi HuangµÈ¡£¡£¡£¡£¡£Lingdong Kong ÊÇ NUS ÅÌËã»úϵ²©Ê¿Éú¡¢ÏîĿǣͷÈË£¬£¬£¬£¬£¬ºã¾Ãά»¤×Ô¶¯¼ÝÊ»ºÍ 3D/4D ÌìϽ¨Ä£Ïà¹ØÏîÄ¿£»£»£»Ziwei Liu ÊÇ NTU MMLab µÄ Associate Professor / Provost's Chair in AI£¬£¬£¬£¬£¬Ñо¿ÁýÕÖÅÌËã»úÊÓ¾õ¡¢»úеѧϰºÍÅÌËã»úͼÐΣ»£»£»Liang Pan ÊÇ Shanghai AI Lab Ñо¿¿Æѧ¼Ò£¬£¬£¬£¬£¬ºã¾Ã×ö 3D ÊÓ¾õ¡¢world models ºÍ embodied AI£»£»£»Wei Tsang Ooi ÊÇ NUS ¶àýϵһÇÐÆ«Ïò½ÌÊÚ¡£¡£¡£¡£¡£Õâ¸ö×éºÏÈà WorldLens ͬʱվÔÚ×Ô¶¯¼ÝÊ»¡¢4D ³¡¾°¡¢3D ÊÓ¾õºÍÈËÀàÆ«ºÃÆÀ²âµÄ½»Ö¯µãÉÏ¡£¡£¡£¡£¡£

ΪʲôֵµÃ¹Ø×¢

×Ô¶¯¼ÝÊ»ÊÇÌìÏÂÄ£×Ó×îÈÝÒ×ÂäµØ¡¢Ò²×î²»¿ÉÖ»¿´ demo µÄÁìÓò¡£¡£¡£¡£¡£WorldLens µÄ¼ÛÖµÔÚÓÚ°Ñ¡°ÌìÉúÌìÏ¡±À­»Øµ½ÁËÇå¾²¡¢±Õ»·ºÍÏÂÓÎʹÃüÉÏ£ºÒ»¸öÄ£×ÓÌìÉúµÄÊÓƵÔÙƯÁÁ£¬£¬£¬£¬£¬ÈôÊÇ»áÈøÐ֪ģ×ÓÍË»¯£¬£¬£¬£¬£¬»òÕßÈà planner ÔÚÄ£ÄâÌìÏÂÀïײ³µ£¬£¬£¬£¬£¬Ëü¾Í²»¿ÉËã¿É¿¿µÄ¼ÝÊ»ÌìÏÂÄ£×Ó¡£¡£¡£¡£¡£

ÏîÄ¿Ò³£ºhttps://worldbench.github.io/worldlens

ÂÛÎÄ£ºhttps://arxiv.org/abs/2512.10958

´úÂ룺https://github.com/worldbench/WorldLens

Leaderboard£ºhttps://huggingface.co/spaces/worldbench/worldlens

4. WorldArena 2.0

Ò»¾ä»°ÏÈÈÝ£º WorldArena ÆÀ¹À¾ßÉíÌìÏÂÄ£×ÓÊÇ·ñÕæµÄÄÜ·þÎñ»úеÈËʹÃü£¬£¬£¬£¬£¬¶ø²»µ«½öÊÇÌìÉúÒ»¶Î¿´ÆðÀ´²»´íµÄ²Ù×÷ÊÓƵ¡£¡£¡£¡£¡£

WorldArena ½â¾öµÄÊǾßÉíÖÇÄÜÀï¡°¸ÐÖªÖÊÁ¿ºÍ¹¦Ð§¼ÛÖµÍѽڡ±µÄÎÊÌâ¡£¡£¡£¡£¡£Ðí¶à embodied world model ¿ÉÒÔÌìÉú¸ßÖÊÁ¿ÊÓƵ£¬£¬£¬£¬£¬µ«»úеÈËÕæÕýÌåÌùµÄÊÇ£ºËüÄܲ»¿ÉºÏ³ÉÓÐÓÃѵÁ·Êý¾Ý£¿£¿£¿£¿£¿£¿Äܲ»¿É×÷Ϊ policy evaluator ÆÀÔ¤Õ½ÂÔ£¿£¿£¿£¿£¿£¿Äܲ»¿É×ÊÖú agent ×ö action planning£¿£¿£¿£¿£¿£¿WorldArena ÏÈÓà 16 ¸öÖ¸±êÆÀ²âÊÓƵ¸ÐÖªÖÊÁ¿£¬£¬£¬£¬£¬ÁýÕÖÊÓ¾õÖÊÁ¿¡¢Ô˶¯ÖÊÁ¿¡¢ÄÚÈÝÒ»ÖÂÐÔ¡¢ÎïÀí×ñÊØ¡¢3D ׼ȷÐÔ¡¢¿É¿ØÐÔ 6 ¸ö×Óά¶È£¬£¬£¬£¬£¬ÔÙÆÀ¹ÀÈýÀà¾ßÉí¹¦Ð§£ºData Engine¡¢Policy Evaluator¡¢Action Planner£¬£¬£¬£¬£¬²¢Ìá³ö EWMScore ×öͳһ»³±§¡£¡£¡£¡£¡£

WorldArena 2.0 ÔÚÕâ¸ö»ù´¡ÉϼÌÐøÀ©Õ¹£ºµÚÒ»£¬£¬£¬£¬£¬Ä£Ì¬´Ó vision-only À©Õ¹µ½ visuo-tactile£¬£¬£¬£¬£¬ÈÃÄ£×Ó±ØÐè´¦Àí½Ó´¥¡¢Á¦¡¢»¬¶¯ºÍÖÊÁϽ»»¥£»£»£»µÚ¶þ£¬£¬£¬£¬£¬¹¦Ð§´ÓÀëÏßÆÀ¹ÀÀ©Õ¹µ½ online RL£¬£¬£¬£¬£¬°ÑÌìÏÂÄ£×Ó¿´³É¿É½»»¥ÇéÐÎÀ´ÑµÁ·ºÍË¢ÐÂÕ½ÂÔ£»£»£»µÚÈý£¬£¬£¬£¬£¬Æ½Ì¨´Ó simulator-only À©Õ¹µ½ RoboTwin 2.0¡¢LIBERO ºÍ AgileX split-type ALOHA Õæʵ»úеÈË¡£¡£¡£¡£¡£ËüµÄ½áÂÛÒ²¸üÌù½ü¹¤ÒµÎÊÌ⣺ģÄâÇéÐÎÀïµÄºÃЧ¹û²»¼´ÊÇÕæʵ°²ÅÅÄÜÁ¦£¬£¬£¬£¬£¬sim-to-real gap ÈÔÈ»ÊǾßÉíÌìÏÂÄ£×Ó±ØÐè¿ç¹ýµÄÓ²Ãż÷¡£¡£¡£¡£¡£

¶à»ú¹¹ÁªºÏÁªºÏÐû²¼£ºTsinghua University¡¢Shanghai Jiao Tong University¡¢The University of Hong Kong¡¢Princeton University¡¢Chinese Academy of Sciences¡¢University of Science and Technology of China¡¢Peking University¡¢National University of Singapore¡£¡£¡£¡£¡£×÷Õß°üÀ¨ Yu Shang¡¢Zhuohang Li¡¢Yiding Ma¡¢Weikang Su¡¢Xin JinµÈ¡£¡£¡£¡£¡£Yong Li ÊÇÇ廪 FIB Lab Ö÷ÒªÈÏÕæÈË£¬£¬£¬£¬£¬ºã¾Ã×ö AI¡¢Êý¾ÝÍÚ¾òºÍ¶¼»áÖÇÄÜ£»£»£»Jun Zhu ÊÇÇ廪 Bosch AI Professor¡¢IEEE/AAAI Fellow£¬£¬£¬£¬£¬Ñо¿»úеѧϰ¡¢Éî¶ÈÌìÉúÄ£×Ó¡¢Ç¿»¯Ñ§Ï°ºÍ·´¿¹Â³°ôÐÔ£»£»£»Wenwu Zhu ÊÇÇ廪ÅÌËã»úϵ½ÌÊÚ£¬£¬£¬£¬£¬ÔøÔÚ Microsoft Research Asia¡¢Intel Research China µÈ¹¤ÒµÑо¿»ú¹¹µ£µ±Ö÷Òª½ÇÉ«£»£»£»Tat-Seng Chua ÊÇ NUS School of Computing µÄ KITHCT Chair Professor£¬£¬£¬£¬£¬Ò²ÔøÈÎ NUS Computing ´´ÔºÔº³¤¡£¡£¡£¡£¡£Õâ¸öÕóÈÝÈà WorldArena ¸üÏñ¾ßÉíÌìÏÂÄ£×ÓÁìÓòµÄ¡°Ñ§Êõ-»úеÈË-¶àýÌå-ϵͳ¡±ÁªºÏÆÀ²â¡£¡£¡£¡£¡£

ΪʲôֵµÃ¹Ø×¢

Ëü°ÑÌìÏÂÄ£×Ó´Ó¡°»á²»»áÌìÉúδÀ´Ö¡¡±Íƽøµ½¡°Äܲ»¿ÉÈûúеÈ˱äÇ¿¡±¡£¡£¡£¡£¡£Õâ¶Ô¾ßÉíÖÇÄܺÜÊÇÒªº¦£¬£¬£¬£¬£¬ÓÉÓÚÕæÕýµÄ»úеÈËÌìÏÂÄ£×Ó²»µ«ÊÇÄ£Äâ»­Ã棬£¬£¬£¬£¬¶øÊÇÒª³ÉΪÊý¾ÝÒýÇæ¡¢ÇéÐÎÊðÀíºÍÍýÏëÄ£¿£¿£¿£¿£¿£¿é¡£¡£¡£¡£¡£WorldArena ÊÇÏÖÔÚÉÙÊý°Ñ perception ºÍ functional utility ·ÅÔÚͳһ¸ö¿ò¼ÜÀïÑÏËàÆÀ¹ÀµÄ benchmark¡£¡£¡£¡£¡£

ÏîÄ¿Ò³£ºhttps://world-arena.ai/

ÂÛÎÄ£ºhttps://arxiv.org/abs/2602.08971

´úÂ룺https://github.com/tsinghua-fib-lab/WorldArena

Leaderboard£ºhttps://huggingface.co/spaces/WorldArena/WorldArena

5. Physics-IQ / Physics-IQ Verified

Ò»¾ä»°ÏÈÈÝ£º Physics-IQ ÓÃÕæʵÎïÀíʵÑéÊÓƵ¿¼ÊÓƵÌìÉúÄ£×Ó£¬£¬£¬£¬£¬Ö±½Ó×·ÎÊÒ»¸ö»ù´¡ÎÊÌ⣺ģ×ÓÊÇÔÚÃ÷È·ÎïÀí£¬£¬£¬£¬£¬ÕÕ¾ÉÖ»ÊÇÔÚÕ¹ÍûÏñËØ£¿£¿£¿£¿£¿£¿

Physics-IQ ½â¾öµÄÊÇ¡°ÊÓ¾õÕæʵ²»¼´ÊÇÎïÀíÃ÷È·¡±µÄÎÊÌâ¡£¡£¡£¡£¡£ËüÒªÇóÄ£×Óƾ֤ÕæʵÌìÏÂÎïÀíÊÓƵµÄÇ°°ë¶ÎÕ¹ÍûºóÐøÑÝ»¯£¬£¬£¬£¬£¬ÁýÕÖ¹ÌÌåÁ¦Ñ§¡¢Á÷Ì嶯Á¦Ñ§¡¢¹âѧ¡¢ÈÈÁ¦Ñ§¡¢´ÅѧÎåÀàÎïÀíÕ÷Ï󡣡£¡£¡£¡£ÆÀ²â²»¿´ÊÓƵÊÇ·ñ¡°¿´ÆðÀ´¸ß¼¶¡±£¬£¬£¬£¬£¬¶øÊÇ¿´Ä£×ÓÕ¹ÍûµÄÎïÀíת±äÔڿռ䡢ʱ¼äºÍÔ˶¯ÇøÓòÉÏÊÇ·ñ¿¿½üÕæʵЧ¹û¡£¡£¡£¡£¡£ÂÛÎÄÆÀ¹ÀÁË Sora¡¢Runway¡¢Pika¡¢Lumiere¡¢Stable Video Diffusion¡¢VideoPoet µÈÄ£×Ó£¬£¬£¬£¬£¬½áÂÛÏ൱ֱ½Ó£ºÄ¿½ñÊÓƵÌìÉúÄ£×ÓµÄÎïÀíÃ÷È·ÈÔÈ»ÓÐÏÞ£¬£¬£¬£¬£¬²¢ÇÒÓëÊÓ¾õÕæʵ¸Ð²¢²»Ç¿Ïà¹Ø£»£»£»¶àÖ¡Ìõ¼þͨ³£±Èµ¥Ö¡Ìõ¼þ¸üÓÐ×ÊÖú£¬£¬£¬£¬£¬Á÷ÌåÕ÷ÏóÏà¶Ô¸üÈÝÒ×£¬£¬£¬£¬£¬¹ÌÌåÁ¦Ñ§¸üÄÑ¡£¡£¡£¡£¡£2026 Äê 6 ÔÂÓÖ·ºÆðÁË Physics-IQ Verified£¬£¬£¬£¬£¬¶ÔԭʼÑù±¾ºÍ prompt ×öÁËϵͳÉó¼ÆºÍÐÞ¶©£»£»£»ÈôÊÇÊÇÏÖʵÅÜÆÀ²â£¬£¬£¬£¬£¬ÓÅÏȹØ×¢ Verified °æ±¾»á¸üÎÈ¡£¡£¡£¡£¡£

Physics-IQ ÓÉ Saman Motamed¡¢Laura Culp¡¢Kevin Swersky¡¢Priyank Jaini¡¢Robert Geirhos Ðû²¼£¬£¬£¬£¬£¬ÂÛÎÄÊ×Ò³±ê×¢ Saman Motamed À´×Ô INSAIT / Sofia University£¬£¬£¬£¬£¬Ïà¹ØÊÂÇéÔÚ Google DeepMind ʱ´úÍê³É£¬£¬£¬£¬£¬ÆäÓà×÷ÕßÀ´×Ô Google DeepMind¡£¡£¡£¡£¡£INSAIT ¹Ù·½±¨µÀÒ²Ã÷È·³Æ¸Ã benchmark ÓÉ INSAIT Óë Google DeepMind ÁªºÏ¿ª·¢£¬£¬£¬£¬£¬²¢ÔÚ ICCV 2025 Òý·¢¹Ø×¢¡£¡£¡£¡£¡£Saman Motamed ÊÇ INSAIT µÄ ELLIS PhD student£¬£¬£¬£¬£¬Ñо¿ÅÌËã»úÊÓ¾õ¡¢ÌìÉúÄ£×ÓºÍ intuitive physics£»£»£»Robert Geirhos ÊÇ Google DeepMind Staff Research Scientist£¬£¬£¬£¬£¬ºã¾Ã×öÊÓ¾õ³°ôÐÔ¡¢¿ÉÚ¹ÊÍÐÔºÍÈË»úÊÓ¾õ½ÏÁ¿£¬£¬£¬£¬£¬´ú±íÊÂÇé°üÀ¨Éî¶ÈÍøÂç texture bias / shortcut learning Æ«Ïò£»£»£»Priyank Jaini ºÍ Kevin Swersky Ò²¶¼À´×Ô Google DeepMind µÄÌìÉúÄ£×ÓÓë»úеѧϰÑо¿Ïß¡£¡£¡£¡£¡£Õâ¸öÍŶӵÄÓÅÊÆÊÇ°ÑÕæʵÎïÀíʵÑé¡¢ÊÓ¾õÐÄÀíѧʽÎÊÌâÒâʶºÍÇ°ÑØÊÓƵģ×ÓÑо¿·Åµ½ÁËÒ»Æ𡣡£¡£¡£¡£

ΪʲôֵµÃ¹Ø×¢

ÎïÀíÊÇÌìÏÂÄ£×ÓÐðÊÂÀï×îÄÑÈÆ¿ªµÄµ×²ãÄÜÁ¦¡£¡£¡£¡£¡£Ò»¸öÄ£×Ó¿ÉÒÔÌìÉúÓ°Ï·¸ÐË®»¨¡¢±¬Õ¨ºÍÅöײ£¬£¬£¬£¬£¬µ«ÈôÊÇËüÎÞ·¨Õ¹ÍûÕæʵÎïÌå»áÔõÑù¼ÌÐøÔ˶¯£¬£¬£¬£¬£¬Ëü¾ÍºÜÄѳÉΪ¿É¿¿µÄÌìÏÂÄ£×Ó¡£¡£¡£¡£¡£Physics-IQ µÄ¼ÛÖµÔÚÓÚËü×ã¹»Ö±½Ó£ºÉÙ½²¡°Ïñ²»Ïñ¡±£¬£¬£¬£¬£¬¶àÎÊ¡°¶Ô²î³Ø¡±¡£¡£¡£¡£¡£

ÏîÄ¿Ò³£ºhttps://physics-iq.github.io/

ÂÛÎÄ£ºhttps://arxiv.org/abs/2501.09038

´úÂ룺https://github.com/google-deepmind/physics-iq-benchmark

Physics-IQ Verified£ºhttps://arxiv.org/abs/2606.18943

ÌìÏÂÄ£×ÓÏÖÔÚ´¦ÔÚÒ»¸ö¿´·¨¿ìËÙÉýΡ¢ÊÖÒÕõ辶·Ö²æ¡¢ÆÀ²âϵͳ׷׏¤ÒµÐðʲ¹¿ÎµÄÔÓÂÒÆÚ¡£¡£¡£¡£¡£Ò²ÕýÓÉÓÚÔÓÂÒ£¬£¬£¬£¬£¬benchmark ²ÅÖµµÃ¿´£ºËüÃÇÊÇÕⳡ½ç˵ÕùÈ¡À£¬£¬£¬£¬ÉÙÊýÄÜ°Ñ¿ÚºÅÂäµ½²âÊÔÌâÉϵŤ¾ß¡£¡£¡£¡£¡£

Benchmark ¿ÉÒÔ¸øÒ»¸ö»¹ºÜËÉÉ¢µÄ¿´·¨»­½çÏß¡£¡£¡£¡£¡£WorldModelBench ¿¼²ìÊÓƵģ×ÓÓÐûÓÐ֪ʶºÍÎïÀíÔ¼Êø£»£»£»WorldScore Ȩºâ²î±ðÌìÏÂÌìÉúõ辶Äܲ»¿Éͳһ½ÏÁ¿£»£»£»WorldLens ¿ÉÒÔÓÃÀ´ÆÀ¹À¼ÝÊ»ÌìÏÂÊÇ·ñÕæµÄ¶ÔÇå¾²ºÍÏÂÓÎʹÃüÓÐÓ㻣»£»WorldArena ¿´µÄÊǾßÉíÌìÏÂÄ£×ÓÊÇ·ñÕæµÄÄÜ·þÎñ»úеÈË£»£»£»Physics-IQ Ôò°Ñ½çÏßѹµ½×îµ×²ã£¬£¬£¬£¬£¬ÎÊÄ£×Ó¾¿¾¹ÊÇÔÚÃ÷È·ÎïÀí£¬£¬£¬£¬£¬ÕÕ¾ÉÖ»ÊÇÔÚ²¹ÏñËØ¡£¡£¡£¡£¡£

ÌìÏÂÄ£×ÓÉú³¤µ½½ñÌ죬£¬£¬£¬£¬ÆÀ²âµÄÖ÷ÒªÐÔÕýÔÚÈÕÒæ͹ÏÔ£¬£¬£¬£¬£¬ÓïÑÔÄ£×ÓÒѾ­¸üÔç֤ʵÎúÕâ¼þÊ£¬£¬£¬£¬£¬½è¼øÓïÑÔÄ£×ÓµÄÂÄÀú£¬£¬£¬£¬£¬ÔçÈÕÈ·¶¨½¹µãÆÀ²âÆ«Ïò£¬£¬£¬£¬£¬Ò²½«³ÉΪδÀ´¾ºÕùµÄÒªº¦¡£¡£¡£¡£¡£

µã¸ö¡°°®ÐÄ¡±£¬£¬£¬£¬£¬ÔÙ×ß °É

±¾ÖÜÑо¿»ú¹¹×ª´ï×îÐÂÐÐҵϣÍû±¾ÖÜ AI ÏîÄ¿ÍƼö£ºWorldModelBench£¬£¬£¬£¬£¬WorldScore£¬£¬£¬£¬£¬WorldLens¡­ÌìÏÂÄ£×Ó¾®Å磬£¬£¬£¬£¬ÐèҪеÄBenchmark

×÷Õߣü·®ÑÅæÃ΢ÐÅ£ü FFF111f__

2026 ÄêÉÏ°ëÄ꣬£¬£¬£¬£¬¡°ÌìÏÂÄ£×Ó¡±ÒѾ­´ÓÂÛÎÄÀïµÄÊÖÒÕ´Ê£¬£¬£¬£¬£¬Äð³ÉÁË AI ¹«Ë¾¡¢»úеÈ˹«Ë¾¡¢×Ô¶¯¼ÝÊ»¹«Ë¾ºÍ VC ÌÖÂÛÀïÈƲ»¿ªµÄÒªº¦´Ê¡£¡£¡£¡£¡£ÊÓƵÌìÉúÄ£×Ó˵×Ô¼ºÄÜÄ£ÄâÎïÀíÌìÏ£¬£¬£¬£¬£¬¾ßÉíÖÇÄÜÍŶÓ˵×Ô¼ºÒªÓÃÌìÏÂÄ£×ÓѵÁ·»úеÈË£¬£¬£¬£¬£¬×Ô¶¯¼ÝÊ»¹«Ë¾ËµÌìÉúʽ·ÂÕæ¿ÉÒÔÌæ»»Ò»²¿·ÖÕæʵ·²â¡£¡£¡£¡£¡£

µ«ÎÊÌâÒ²ËæÖ®±äµÃ¼âÈñ£ºÒ»¸öÄ£×Ó»áÌìÉúƯÁÁÊÓƵ£¬£¬£¬£¬£¬¾Í¼´ÊÇËüÃ÷È·ÌìÏÂÁËÂ𣿣¿£¿£¿£¿£¿ËüÄܲ»¿É×ñÊØÎïÀí¼ÍÂÉ£¿£¿£¿£¿£¿£¿Äܲ»¿É¼á³Ö 3D ¼¸ºÎÒ»Ö£¿£¿£¿£¿£¿£¿Äܲ»¿É·þÎñ×Ô¶¯¼ÝÊ»ºÍ»úеÈ˾öÒ飿£¿£¿£¿£¿£¿Äܲ»¿É´Ó¡°¿´ÆðÀ´ÏñÕæµÄ¡±£¬£¬£¬£¬£¬×ßµ½¡°ÓÃÆðÀ´ÕæµÄÓÐÓá±£¿£¿£¿£¿£¿£¿

ÒÔÊÇ£¬£¬£¬£¬£¬±ÈÆðÖ»¿´ demo£¬£¬£¬£¬£¬¸üÓ¦¸Ã¿´ benchmark¡£¡£¡£¡£¡£ÏÂÃæÕâÎå¸öÌìÏÂÄ£×Óbenchmark°üÀ¨Í¨ÓÃÊÓƵÌìÏÂÄ£×Ó¡¢Í³Ò»ÌìÏÂÌìÉú¡¢×Ô¶¯¼ÝÊ»¡¢¾ßÉíÖÇÄÜ¡¢ÎïÀíÍÆÀí¡£¡£¡£¡£¡£ËüÃDz»ÊÇËùÓÐ benchmark µÄÈ«¼¯£¬£¬£¬£¬£¬µ«»ù±¾ÁýÕÖÁ˵±ÏÂÌìÏÂÄ£×Ó×îÒªº¦µÄÎåÌõÖ÷Ïß¡£¡£¡£¡£¡£

1. WorldModelBench

Ò»¾ä»°ÏÈÈÝ£º WorldModelBench °ÑÊÓƵÌìÉúÄ£×Ó¿´³ÉÌìÏÂÄ£×ÓÀ´¿¼£¬£¬£¬£¬£¬²»µ«ÎÊ»­ÃæºÃÇ·ÔÃÄ¿£¬£¬£¬£¬£¬»¹ÎÊÖ¸Á֪ʶºÍÎïÀíÊÇ·ñ¾­µÃÆð¼ì²é¡£¡£¡£¡£¡£

WorldModelBench ½â¾öµÄÊÇͨÓÃÊÓƵģ×ÓÆÀ²âÀï×î½¹µãµÄȱ¿Ú£ºÒÑÍùÐí¶à benchmark Æ«ÏòÊÓ¾õÖÊÁ¿¡¢Îı¾¶ÔÆëijÈËÀàÆ«ºÃ£¬£¬£¬£¬£¬µ«¡°ÌìÏÂÄ£×Ó¡±ÕæÕýÒª»Ø¸²µÄÊÇ£¬£¬£¬£¬£¬Ä£×ÓÊÇ·ñÃ÷È·Ðж¯¡¢³¡¾°Òò¹ûºÍ»ù´¡ÎïÀí¡£¡£¡£¡£¡£ËüÁýÕÖ»úеÈË¡¢×Ô¶¯¼ÝÊ»¡¢¹¤Òµ¡¢ÈËÀà»î¶¯¡¢ÓÎÏ·¡¢¶¯»­¡¢×ÔÈ»µÈ 7 ¸öÓ¦ÓÃÓò£¬£¬£¬£¬£¬Ï¸·Öµ½ 56 ¸ö×ÓÓò£¬£¬£¬£¬£¬Ê¹Óà 350 ¸öͼÎÄÌáÐÑ£¬£¬£¬£¬£¬²¢ÓÃÖ¸Áî×·Ë桢֪ʶ¡¢ÎïÀí×ñÊØÈý´óά¶ÈÆÀ²â¡£¡£¡£¡£¡£ÎïÀí²¿·ÖÖص㿴ţ¶ÙµÚÒ»¶¨ÂÉ¡¢Ðα䡢Á÷Ìå¡¢´©Í¸¡¢ÖØÁ¦µÈ³£¼û¡°ÌìϽ¨Ä£»£»£»Ã¾õ¡±¡£¡£¡£¡£¡£ÂÛÎÄ»¹ÖÚ°üÁË 6.7 Íò¸öÈËÀà±ê×¢£¬£¬£¬£¬£¬²¢ÑµÁ·ÁËÒ»¸ö human-aligned judge À´×Ô¶¯ÆÀ²â¡£¡£¡£¡£¡£ÔÚЧ¹ûÉÏ£¬£¬£¬£¬£¬Kling µÈ±ÕÔ´Ä£×ÓÕûÌå¿¿Ç°£¬£¬£¬£¬£¬Mochi¡¢OpenSoraPlan µÈ¿ªÔ´Ä£×ÓÔÚ²¿·Öά¶È¿¿½ü±ÕÔ´Ä£×Ó£¬£¬£¬£¬£¬µ«ÂÛÎĵĽ¹µã½áÂÛ²»ÊÇ¡°Ë­ÒѾ­½â¾öÁËÌìÏÂÄ£×Ó¡±£¬£¬£¬£¬£¬¶øÊÇÍ·²¿Ä£×ÓÈÔÈ»Àë¿É¿¿ÌìÏÂÄ£ÄâÆ÷ºÜÔ¶¡£¡£¡£¡£¡£

Ðû²¼ÍŶÓÀ´×Ô UC Berkeley¡¢UC San Diego¡¢NVIDIA ºÍ MIT £¬£¬£¬£¬£¬Ion Stoica ÊÇ UC Berkeley ÅÌËã»ú½ÌÊÚ£¬£¬£¬£¬£¬Ò²ÊÇ Databricks¡¢Anyscale µÈ¹«Ë¾µÄÅäºÏÊ×´´ÈË£»£»£»Joseph E. Gonzalez ÊÇ Berkeley EECS ½ÌÊÚ¡¢Sky Computing Lab / RISE Lab ÅäºÏÈÏÕæÈË£¬£¬£¬£¬£¬²¢ÊôÓÚ BAIR Éú̬¡£¡£¡£¡£¡£Õâʹ WorldModelBench ²»µ«ÊÇÒ»¸öÊÓƵÆÀ·Ö¼¯£¬£¬£¬£¬£¬¶øÊÇϵͳ¡¢ÊÓ¾õ¡¢»úеÈ˺͸ßЧ AI ÍŶÓÅäºÏÍƶ¯µÄÌìÏÂÄ£×ÓÆÀ²â¡£¡£¡£¡£¡£

ΪʲôֵµÃ¹Ø×¢

ËüÊÇ×îÊÊÏàÖúΪ¡°Í¨ÓÃÊÓƵÌìÏÂÄ£×ÓÈëÃÅ»ù×¼¡±µÄ benchmark¡£¡£¡£¡£¡£ÓÉÓÚËüûÓÐֻͣÁôÔÚÉóÃÀºÍÇåÎú¶È£¬£¬£¬£¬£¬¶øÊÇ°ÑÌìÏÂÄ£×Ó²ð³ÉÓ¦ÓÃÓò¡¢ÎïÀí¼ÍÂɺÍϸÁ£¶È¹ýʧÀàÐÍ£¬£¬£¬£¬£¬ÌØÊâÊʺÏÅжÏÒ»¸öÊÓƵģ×Ó¾¿¾¹ÊÇÔÚ¡°ÌìÉúÏñÌìϵÄͼÏñ¡±£¬£¬£¬£¬£¬ÕÕ¾ÉÔÚ¡°Ä£ÄâÒ»¸öÄÜ×ÔÇ¢ÑÝ»¯µÄÌìÏ¡±¡£¡£¡£¡£¡£

ÏîÄ¿£ºhttps://worldmodelbench-team.github.io/

ÂÛÎÄ£ºhttps://arxiv.org/abs/2502.20694

´úÂëÓëÆÀ²â£ºhttps://github.com/WorldModelBench-Team/WorldModelBench

²âÊÔÊý¾Ý£ºhttps://huggingface.co/WorldModelBench-Team

2. WorldScore

Ò»¾ä»°ÏÈÈÝ£º WorldScore °Ñ 3D¡¢4D¡¢Í¼ÉúÊÓƵ¡¢ÎÄÉúÊÓƵģ×ӷŵ½Í³Ò»¸ö¡°ÏÂÒ»³¡¾°ÌìÉú¡±¿ò¼ÜÀïÆÀ²â£¬£¬£¬£¬£¬Öص㿴ģ×ÓÄܲ»¿É°´Ïà»ú¹ì¼£Ò»Á¬ÌìÉúÌìÏ¡£¡£¡£¡£¡£

¸ÅÊö£º WorldScore ½â¾öµÄÊÇ¡°²î±ðÌìÏÂÌìÉúõ辶ÎÞ·¨¹«Õý½ÏÁ¿¡±µÄÎÊÌâ¡£¡£¡£¡£¡£3D Ä£×Ó¡¢4D Ä£×ÓºÍÊÓƵģ×ÓÊä³öÐÎ̬²î±ð£¬£¬£¬£¬£¬ÒÑÍùºÜÄÑÔÚͳһÕÅÄÚÍâ½ÏÁ¿¡£¡£¡£¡£¡£WorldScore ½«ÌìÏÂÌìÉú²ð³ÉһϵÁлùÓÚÃ÷È·Ïà»ú¹ì¼£µÄ next-scene generation ʹÃü£¬£¬£¬£¬£¬Óà 3000 ¸öÑùÀýÁýÕÖ¾²Ì¬/¶¯Ì¬¡¢ÊÒÄÚ/ÊÒÍ⡢дʵ/Æø¸Å»¯µÈ³¡¾°£¬£¬£¬£¬£¬²¢´Ó¿É¿ØÐÔ¡¢ÖÊÁ¿¡¢¶¯Ì¬Èý·½Ãæ´ò·Ö¡£¡£¡£¡£¡£¿£¿£¿£¿£¿£¿É¿ØÐÔ¿´Ïà»ú¿ØÖÆ¡¢ÎïÌå¿ØÖƺÍÄÚÈݶÔÆ룬£¬£¬£¬£¬ÖÊÁ¿¿´ 3D Ò»ÖÂÐÔ¡¢¹â¶ÈÒ»ÖÂÐÔ¡¢Æø¸ÅÒ»ÖÂÐÔºÍÖ÷¹ÛÖÊÁ¿£¬£¬£¬£¬£¬¶¯Ì¬¿´Ô˶¯×¼È·ÐÔ¡¢·ù¶ÈºÍÁ÷ͨ¶È¡£¡£¡£¡£¡£¹Ù·½Ð§¹ûÏÔʾ£¬£¬£¬£¬£¬Voyager¡¢WonderWorld¡¢LucidDreamer µÈ 3D/³¡¾°ÌìÉúÄ£×ÓÔÚ¾²Ì¬ÌìÏÂÌìÉúÉϺÜÇ¿£¬£¬£¬£¬£¬CogVideoX-I2V µÈÊÓƵģ×ÓÔÚ¶¯Ì¬·ÖÊýÉÏÌåÏÖÍ»³ö£»£»£»Í¬Ê±£¬£¬£¬£¬£¬T2V Ä£×Óͨ³£¸üÈÝÒ×±»Îı¾¿ØÖÆ£¬£¬£¬£¬£¬I2V Ä£×ÓÍùÍù»­ÖʸüÎÈ¡£¡£¡£¡£¡£

WorldScore ÊÇ Stanford University ÍŶÓÐû²¼£¬£¬£¬£¬£¬ÂÛÎÄ×÷ÕßΪ Haoyi Duan¡¢Hong-Xing Yu¡¢Sirui Chen¡¢Li Fei-Fei¡¢Jiajun Wu¡£¡£¡£¡£¡£Li Fei-Fei ÊÇ Stanford ÅÌËã»úϵ Sequoia Professor¡¢Stanford HAI Ê×´´ÁªºÏÖ÷ÈΣ¬£¬£¬£¬£¬ÔøÈÎ Stanford AI Lab Ö÷ÈÎºÍ Google Cloud AI/ML Ê×ϯ¿Æѧ¼Ò£¬£¬£¬£¬£¬¿ª°ìWorld Labs£¬£¬£¬£¬£¬Ò²ÊÇ ImageNet µÄÒªº¦Íƶ¯Õߣ»£»£»Jiajun Wu ÊÇ Stanford ÅÌËã»úϵÖúÀí½ÌÊÚ£¬£¬£¬£¬£¬Í¬Ê±¼æÈÎÐÄÀíѧƫÏò£¬£¬£¬£¬£¬Ñо¿ÅÌËã»úÊÓ¾õ¡¢»úеÈ˺ÍÅÌËãÈÏÖª¿Æѧ£¬£¬£¬£¬£¬ËùÔÚÆ«Ïò×ÔÈ»Ìù½ü¡°ÌìÏÂÔõÑù±»ÌåÏÖ¡¢ÍÆÀíºÍÌìÉú¡±¡£¡£¡£¡£¡£Òò´Ë£¬£¬£¬£¬£¬WorldScore ¸üÏñÊÇ Stanford ÊÓ¾õѧϰ¡¢3D/ÎïÀí³¡¾°Ã÷È·ºÍÌìÉúÄ£×ÓÑо¿ÂöÂçϵÄÒ»Ì×ÌìÏÂÌìÉú»³±§¡£¡£¡£¡£¡£

ΪʲôֵµÃ¹Ø×¢£º

ÈôÊÇ˵ WorldModelBench ¸üÏñ¡°ÊÓƵģ×ÓÊDz»ÊÇÌìÏÂÄ£×Ó¡±µÄÌå¼ì£¬£¬£¬£¬£¬WorldScore Ôò¸üÏñ¡°ÌìÏÂÌìÉúõ辶´óºáÆÀ¡±¡£¡£¡£¡£¡£Ëü°Ñ 3D¡¢4D¡¢ÊÓƵÈýÌõÊÖÒÕõ辶·Å½øͳһЭÒ飬£¬£¬£¬£¬¶Ô´´Òµ¹«Ë¾ºÍÑо¿ÍŶÓÅжÏÊÖÒÕõ辶ºÜÓмÛÖµ£º¾¿¾¹¸ÃѺÊÓƵÌìÉú¡¢3D ÌìÉú£¬£¬£¬£¬£¬Õվɸü»ìÏýµÄ 4D ÌìÏÂÌìÉú¡£¡£¡£¡£¡£

ÏîÄ¿Ò³£ºhttps://haoyi-duan.github.io/WorldScore/

ÂÛÎÄ£ºhttps://arxiv.org/abs/2504.00983

´úÂ룺https://github.com/haoyi-duan/WorldScore

Êý¾Ý¼¯£ºhttps://huggingface.co/datasets/haoyi-duan/WorldScore

3. WorldLens

Ò»¾ä»°ÏÈÈÝ£º WorldLens רÃÅÆÀ¹À×Ô¶¯¼ÝÊ»ÌìÏÂÄ£×Ó£¬£¬£¬£¬£¬½¹µãÎÊÌâÊÇ£ºÌìÉúµÄ¼ÝÊ»Ììϲ»µ«ÒªÏñ£¬£¬£¬£¬£¬»¹Òª¼¸ºÎÎȹ̡¢ÎïÀíºÏÀí¡¢ÐÐΪÇå¾²¡¢¶ÔÏÂÓÎʹÃüÓÐÓᣡ£¡£¡£¡£

WorldLens ½â¾öµÄÊÇ×Ô¶¯¼ÝÊ»³¡¾°Àï¡°ÊÓ¾õ±ÆÕ浫¹¦Ð§Ê§Õ桱µÄÎÊÌâ¡£¡£¡£¡£¡£¼ÝÊ»ÌìÏÂÄ£×ÓºÜÈÝÒ×ÌìÉú¿´ÆðÀ´ºÏÀíµÄ 4D ³¡¾°£¬£¬£¬£¬£¬µ«Ò»µ©Çе½ÖØÐÞ¡¢±Õ»·¿ØÖÆ¡¢ÏÂÓμì²âÖ§½â£¬£¬£¬£¬£¬¾Í»á̻¶¼¸ºÎƯÒÆ¡¢Ô˶¯²»Îȹ̡¢Åöײ¡¢Æ«Àëõ辶µÈÎÊÌâ¡£¡£¡£¡£¡£WorldLens ÓÃÎå¸ö»¥²¹Æ«ÏòÆÀ²â£ºGeneration¡¢Reconstruction¡¢Action-Following¡¢Downstream Task¡¢Human Preference£¬£¬£¬£¬£¬²¢½øÒ»²½¹¹½¨ WorldLens-26K ÈËÀàÆ«ºÃÊý¾ÝºÍ WorldLens-Agent ÊÓ¾õÓïÑÔÆÀÉóÆ÷¡£¡£¡£¡£¡£ÏîÄ¿½áÂÛºÜÊÇÏÖʵ£ºÃ»Óмòµ¥Ä£×ÓÄÜÔÚËùÓÐά¶ÈͳÖΰñµ¥£¬£¬£¬£¬£¬ÎÆÀíÕæʵµÄÄ£×Ó¿ÉÄÜÎïÀíºÍ¼¸ºÎ²»ÎÈ£¬£¬£¬£¬£¬¼¸ºÎÎȹ̵ÄÄ£×ÓÓÖ¿ÉÄÜÐÐΪϸ½Úȱ·¦¡£¡£¡£¡£¡£ÂÛÎÄʵÑéÖУ¬£¬£¬£¬£¬DiST-4D ÔÚÐÂÊÓ½ÇÖÊÁ¿ºÍ²¿·ÖÏÂÓÎʹÃüÉÏÌåÏÖÇ¿£¬£¬£¬£¬£¬OpenDWM µÈÄ£×ÓÔÚ¼¸ºÎÎȹÌÐÔÉÏÓÐÓÅÊÆ£¬£¬£¬£¬£¬DriveDreamer-2 Ò²ÔÚʹÃüÏà¹ØÖ¸±êÖнøÈëÇ¿ÊÆÌݶÓ¡£¡£¡£¡£¡£

WorldLens ÓÉ WorldBench Team Ðû²¼£¬£¬£¬£¬£¬NUS¡¢University of Macau¡¢USTC¡¢Zhejiang University¡¢NTU¡¢Horizon Robotics¡¢HUST¡¢TUM¡¢Fudan University¡¢Shanghai AI Lab¡¢A*STAR¡¢CNRS µÈ¡£¡£¡£¡£¡£½¹µã×÷Õß°üÀ¨ Ao Liang¡¢Lingdong Kong¡¢Tianyi Yan¡¢Hongsi Liu¡¢Ziqi HuangµÈ¡£¡£¡£¡£¡£Lingdong Kong ÊÇ NUS ÅÌËã»úϵ²©Ê¿Éú¡¢ÏîĿǣͷÈË£¬£¬£¬£¬£¬ºã¾Ãά»¤×Ô¶¯¼ÝÊ»ºÍ 3D/4D ÌìϽ¨Ä£Ïà¹ØÏîÄ¿£»£»£»Ziwei Liu ÊÇ NTU MMLab µÄ Associate Professor / Provost's Chair in AI£¬£¬£¬£¬£¬Ñо¿ÁýÕÖÅÌËã»úÊÓ¾õ¡¢»úеѧϰºÍÅÌËã»úͼÐΣ»£»£»Liang Pan ÊÇ Shanghai AI Lab Ñо¿¿Æѧ¼Ò£¬£¬£¬£¬£¬ºã¾Ã×ö 3D ÊÓ¾õ¡¢world models ºÍ embodied AI£»£»£»Wei Tsang Ooi ÊÇ NUS ¶àýϵһÇÐÆ«Ïò½ÌÊÚ¡£¡£¡£¡£¡£Õâ¸ö×éºÏÈà WorldLens ͬʱվÔÚ×Ô¶¯¼ÝÊ»¡¢4D ³¡¾°¡¢3D ÊÓ¾õºÍÈËÀàÆ«ºÃÆÀ²âµÄ½»Ö¯µãÉÏ¡£¡£¡£¡£¡£

ΪʲôֵµÃ¹Ø×¢

×Ô¶¯¼ÝÊ»ÊÇÌìÏÂÄ£×Ó×îÈÝÒ×ÂäµØ¡¢Ò²×î²»¿ÉÖ»¿´ demo µÄÁìÓò¡£¡£¡£¡£¡£WorldLens µÄ¼ÛÖµÔÚÓÚ°Ñ¡°ÌìÉúÌìÏ¡±À­»Øµ½ÁËÇå¾²¡¢±Õ»·ºÍÏÂÓÎʹÃüÉÏ£ºÒ»¸öÄ£×ÓÌìÉúµÄÊÓƵÔÙƯÁÁ£¬£¬£¬£¬£¬ÈôÊÇ»áÈøÐ֪ģ×ÓÍË»¯£¬£¬£¬£¬£¬»òÕßÈà planner ÔÚÄ£ÄâÌìÏÂÀïײ³µ£¬£¬£¬£¬£¬Ëü¾Í²»¿ÉËã¿É¿¿µÄ¼ÝÊ»ÌìÏÂÄ£×Ó¡£¡£¡£¡£¡£

ÏîÄ¿Ò³£ºhttps://worldbench.github.io/worldlens

ÂÛÎÄ£ºhttps://arxiv.org/abs/2512.10958

´úÂ룺https://github.com/worldbench/WorldLens

Leaderboard£ºhttps://huggingface.co/spaces/worldbench/worldlens

4. WorldArena 2.0

Ò»¾ä»°ÏÈÈÝ£º WorldArena ÆÀ¹À¾ßÉíÌìÏÂÄ£×ÓÊÇ·ñÕæµÄÄÜ·þÎñ»úеÈËʹÃü£¬£¬£¬£¬£¬¶ø²»µ«½öÊÇÌìÉúÒ»¶Î¿´ÆðÀ´²»´íµÄ²Ù×÷ÊÓƵ¡£¡£¡£¡£¡£

WorldArena ½â¾öµÄÊǾßÉíÖÇÄÜÀï¡°¸ÐÖªÖÊÁ¿ºÍ¹¦Ð§¼ÛÖµÍѽڡ±µÄÎÊÌâ¡£¡£¡£¡£¡£Ðí¶à embodied world model ¿ÉÒÔÌìÉú¸ßÖÊÁ¿ÊÓƵ£¬£¬£¬£¬£¬µ«»úеÈËÕæÕýÌåÌùµÄÊÇ£ºËüÄܲ»¿ÉºÏ³ÉÓÐÓÃѵÁ·Êý¾Ý£¿£¿£¿£¿£¿£¿Äܲ»¿É×÷Ϊ policy evaluator ÆÀÔ¤Õ½ÂÔ£¿£¿£¿£¿£¿£¿Äܲ»¿É×ÊÖú agent ×ö action planning£¿£¿£¿£¿£¿£¿WorldArena ÏÈÓà 16 ¸öÖ¸±êÆÀ²âÊÓƵ¸ÐÖªÖÊÁ¿£¬£¬£¬£¬£¬ÁýÕÖÊÓ¾õÖÊÁ¿¡¢Ô˶¯ÖÊÁ¿¡¢ÄÚÈÝÒ»ÖÂÐÔ¡¢ÎïÀí×ñÊØ¡¢3D ׼ȷÐÔ¡¢¿É¿ØÐÔ 6 ¸ö×Óά¶È£¬£¬£¬£¬£¬ÔÙÆÀ¹ÀÈýÀà¾ßÉí¹¦Ð§£ºData Engine¡¢Policy Evaluator¡¢Action Planner£¬£¬£¬£¬£¬²¢Ìá³ö EWMScore ×öͳһ»³±§¡£¡£¡£¡£¡£

WorldArena 2.0 ÔÚÕâ¸ö»ù´¡ÉϼÌÐøÀ©Õ¹£ºµÚÒ»£¬£¬£¬£¬£¬Ä£Ì¬´Ó vision-only À©Õ¹µ½ visuo-tactile£¬£¬£¬£¬£¬ÈÃÄ£×Ó±ØÐè´¦Àí½Ó´¥¡¢Á¦¡¢»¬¶¯ºÍÖÊÁϽ»»¥£»£»£»µÚ¶þ£¬£¬£¬£¬£¬¹¦Ð§´ÓÀëÏßÆÀ¹ÀÀ©Õ¹µ½ online RL£¬£¬£¬£¬£¬°ÑÌìÏÂÄ£×Ó¿´³É¿É½»»¥ÇéÐÎÀ´ÑµÁ·ºÍË¢ÐÂÕ½ÂÔ£»£»£»µÚÈý£¬£¬£¬£¬£¬Æ½Ì¨´Ó simulator-only À©Õ¹µ½ RoboTwin 2.0¡¢LIBERO ºÍ AgileX split-type ALOHA Õæʵ»úеÈË¡£¡£¡£¡£¡£ËüµÄ½áÂÛÒ²¸üÌù½ü¹¤ÒµÎÊÌ⣺ģÄâÇéÐÎÀïµÄºÃЧ¹û²»¼´ÊÇÕæʵ°²ÅÅÄÜÁ¦£¬£¬£¬£¬£¬sim-to-real gap ÈÔÈ»ÊǾßÉíÌìÏÂÄ£×Ó±ØÐè¿ç¹ýµÄÓ²Ãż÷¡£¡£¡£¡£¡£

¶à»ú¹¹ÁªºÏÁªºÏÐû²¼£ºTsinghua University¡¢Shanghai Jiao Tong University¡¢The University of Hong Kong¡¢Princeton University¡¢Chinese Academy of Sciences¡¢University of Science and Technology of China¡¢Peking University¡¢National University of Singapore¡£¡£¡£¡£¡£×÷Õß°üÀ¨ Yu Shang¡¢Zhuohang Li¡¢Yiding Ma¡¢Weikang Su¡¢Xin JinµÈ¡£¡£¡£¡£¡£Yong Li ÊÇÇ廪 FIB Lab Ö÷ÒªÈÏÕæÈË£¬£¬£¬£¬£¬ºã¾Ã×ö AI¡¢Êý¾ÝÍÚ¾òºÍ¶¼»áÖÇÄÜ£»£»£»Jun Zhu ÊÇÇ廪 Bosch AI Professor¡¢IEEE/AAAI Fellow£¬£¬£¬£¬£¬Ñо¿»úеѧϰ¡¢Éî¶ÈÌìÉúÄ£×Ó¡¢Ç¿»¯Ñ§Ï°ºÍ·´¿¹Â³°ôÐÔ£»£»£»Wenwu Zhu ÊÇÇ廪ÅÌËã»úϵ½ÌÊÚ£¬£¬£¬£¬£¬ÔøÔÚ Microsoft Research Asia¡¢Intel Research China µÈ¹¤ÒµÑо¿»ú¹¹µ£µ±Ö÷Òª½ÇÉ«£»£»£»Tat-Seng Chua ÊÇ NUS School of Computing µÄ KITHCT Chair Professor£¬£¬£¬£¬£¬Ò²ÔøÈÎ NUS Computing ´´ÔºÔº³¤¡£¡£¡£¡£¡£Õâ¸öÕóÈÝÈà WorldArena ¸üÏñ¾ßÉíÌìÏÂÄ£×ÓÁìÓòµÄ¡°Ñ§Êõ-»úеÈË-¶àýÌå-ϵͳ¡±ÁªºÏÆÀ²â¡£¡£¡£¡£¡£

ΪʲôֵµÃ¹Ø×¢

Ëü°ÑÌìÏÂÄ£×Ó´Ó¡°»á²»»áÌìÉúδÀ´Ö¡¡±Íƽøµ½¡°Äܲ»¿ÉÈûúеÈ˱äÇ¿¡±¡£¡£¡£¡£¡£Õâ¶Ô¾ßÉíÖÇÄܺÜÊÇÒªº¦£¬£¬£¬£¬£¬ÓÉÓÚÕæÕýµÄ»úеÈËÌìÏÂÄ£×Ó²»µ«ÊÇÄ£Äâ»­Ã棬£¬£¬£¬£¬¶øÊÇÒª³ÉΪÊý¾ÝÒýÇæ¡¢ÇéÐÎÊðÀíºÍÍýÏëÄ£¿£¿£¿£¿£¿£¿é¡£¡£¡£¡£¡£WorldArena ÊÇÏÖÔÚÉÙÊý°Ñ perception ºÍ functional utility ·ÅÔÚͳһ¸ö¿ò¼ÜÀïÑÏËàÆÀ¹ÀµÄ benchmark¡£¡£¡£¡£¡£

ÏîÄ¿Ò³£ºhttps://world-arena.ai/

ÂÛÎÄ£ºhttps://arxiv.org/abs/2602.08971

´úÂ룺https://github.com/tsinghua-fib-lab/WorldArena

Leaderboard£ºhttps://huggingface.co/spaces/WorldArena/WorldArena

5. Physics-IQ / Physics-IQ Verified

Ò»¾ä»°ÏÈÈÝ£º Physics-IQ ÓÃÕæʵÎïÀíʵÑéÊÓƵ¿¼ÊÓƵÌìÉúÄ£×Ó£¬£¬£¬£¬£¬Ö±½Ó×·ÎÊÒ»¸ö»ù´¡ÎÊÌ⣺ģ×ÓÊÇÔÚÃ÷È·ÎïÀí£¬£¬£¬£¬£¬ÕÕ¾ÉÖ»ÊÇÔÚÕ¹ÍûÏñËØ£¿£¿£¿£¿£¿£¿

Physics-IQ ½â¾öµÄÊÇ¡°ÊÓ¾õÕæʵ²»¼´ÊÇÎïÀíÃ÷È·¡±µÄÎÊÌâ¡£¡£¡£¡£¡£ËüÒªÇóÄ£×Óƾ֤ÕæʵÌìÏÂÎïÀíÊÓƵµÄÇ°°ë¶ÎÕ¹ÍûºóÐøÑÝ»¯£¬£¬£¬£¬£¬ÁýÕÖ¹ÌÌåÁ¦Ñ§¡¢Á÷Ì嶯Á¦Ñ§¡¢¹âѧ¡¢ÈÈÁ¦Ñ§¡¢´ÅѧÎåÀàÎïÀíÕ÷Ï󡣡£¡£¡£¡£ÆÀ²â²»¿´ÊÓƵÊÇ·ñ¡°¿´ÆðÀ´¸ß¼¶¡±£¬£¬£¬£¬£¬¶øÊÇ¿´Ä£×ÓÕ¹ÍûµÄÎïÀíת±äÔڿռ䡢ʱ¼äºÍÔ˶¯ÇøÓòÉÏÊÇ·ñ¿¿½üÕæʵЧ¹û¡£¡£¡£¡£¡£ÂÛÎÄÆÀ¹ÀÁË Sora¡¢Runway¡¢Pika¡¢Lumiere¡¢Stable Video Diffusion¡¢VideoPoet µÈÄ£×Ó£¬£¬£¬£¬£¬½áÂÛÏ൱ֱ½Ó£ºÄ¿½ñÊÓƵÌìÉúÄ£×ÓµÄÎïÀíÃ÷È·ÈÔÈ»ÓÐÏÞ£¬£¬£¬£¬£¬²¢ÇÒÓëÊÓ¾õÕæʵ¸Ð²¢²»Ç¿Ïà¹Ø£»£»£»¶àÖ¡Ìõ¼þͨ³£±Èµ¥Ö¡Ìõ¼þ¸üÓÐ×ÊÖú£¬£¬£¬£¬£¬Á÷ÌåÕ÷ÏóÏà¶Ô¸üÈÝÒ×£¬£¬£¬£¬£¬¹ÌÌåÁ¦Ñ§¸üÄÑ¡£¡£¡£¡£¡£2026 Äê 6 ÔÂÓÖ·ºÆðÁË Physics-IQ Verified£¬£¬£¬£¬£¬¶ÔԭʼÑù±¾ºÍ prompt ×öÁËϵͳÉó¼ÆºÍÐÞ¶©£»£»£»ÈôÊÇÊÇÏÖʵÅÜÆÀ²â£¬£¬£¬£¬£¬ÓÅÏȹØ×¢ Verified °æ±¾»á¸üÎÈ¡£¡£¡£¡£¡£

Physics-IQ ÓÉ Saman Motamed¡¢Laura Culp¡¢Kevin Swersky¡¢Priyank Jaini¡¢Robert Geirhos Ðû²¼£¬£¬£¬£¬£¬ÂÛÎÄÊ×Ò³±ê×¢ Saman Motamed À´×Ô INSAIT / Sofia University£¬£¬£¬£¬£¬Ïà¹ØÊÂÇéÔÚ Google DeepMind ʱ´úÍê³É£¬£¬£¬£¬£¬ÆäÓà×÷ÕßÀ´×Ô Google DeepMind¡£¡£¡£¡£¡£INSAIT ¹Ù·½±¨µÀÒ²Ã÷È·³Æ¸Ã benchmark ÓÉ INSAIT Óë Google DeepMind ÁªºÏ¿ª·¢£¬£¬£¬£¬£¬²¢ÔÚ ICCV 2025 Òý·¢¹Ø×¢¡£¡£¡£¡£¡£Saman Motamed ÊÇ INSAIT µÄ ELLIS PhD student£¬£¬£¬£¬£¬Ñо¿ÅÌËã»úÊÓ¾õ¡¢ÌìÉúÄ£×ÓºÍ intuitive physics£»£»£»Robert Geirhos ÊÇ Google DeepMind Staff Research Scientist£¬£¬£¬£¬£¬ºã¾Ã×öÊÓ¾õ³°ôÐÔ¡¢¿ÉÚ¹ÊÍÐÔºÍÈË»úÊÓ¾õ½ÏÁ¿£¬£¬£¬£¬£¬´ú±íÊÂÇé°üÀ¨Éî¶ÈÍøÂç texture bias / shortcut learning Æ«Ïò£»£»£»Priyank Jaini ºÍ Kevin Swersky Ò²¶¼À´×Ô Google DeepMind µÄÌìÉúÄ£×ÓÓë»úеѧϰÑо¿Ïß¡£¡£¡£¡£¡£Õâ¸öÍŶӵÄÓÅÊÆÊÇ°ÑÕæʵÎïÀíʵÑé¡¢ÊÓ¾õÐÄÀíѧʽÎÊÌâÒâʶºÍÇ°ÑØÊÓƵģ×ÓÑо¿·Åµ½ÁËÒ»Æ𡣡£¡£¡£¡£

ΪʲôֵµÃ¹Ø×¢

ÎïÀíÊÇÌìÏÂÄ£×ÓÐðÊÂÀï×îÄÑÈÆ¿ªµÄµ×²ãÄÜÁ¦¡£¡£¡£¡£¡£Ò»¸öÄ£×Ó¿ÉÒÔÌìÉúÓ°Ï·¸ÐË®»¨¡¢±¬Õ¨ºÍÅöײ£¬£¬£¬£¬£¬µ«ÈôÊÇËüÎÞ·¨Õ¹ÍûÕæʵÎïÌå»áÔõÑù¼ÌÐøÔ˶¯£¬£¬£¬£¬£¬Ëü¾ÍºÜÄѳÉΪ¿É¿¿µÄÌìÏÂÄ£×Ó¡£¡£¡£¡£¡£Physics-IQ µÄ¼ÛÖµÔÚÓÚËü×ã¹»Ö±½Ó£ºÉÙ½²¡°Ïñ²»Ïñ¡±£¬£¬£¬£¬£¬¶àÎÊ¡°¶Ô²î³Ø¡±¡£¡£¡£¡£¡£

ÏîÄ¿Ò³£ºhttps://physics-iq.github.io/

ÂÛÎÄ£ºhttps://arxiv.org/abs/2501.09038

´úÂ룺https://github.com/google-deepmind/physics-iq-benchmark

Physics-IQ Verified£ºhttps://arxiv.org/abs/2606.18943

ÌìÏÂÄ£×ÓÏÖÔÚ´¦ÔÚÒ»¸ö¿´·¨¿ìËÙÉýΡ¢ÊÖÒÕõ辶·Ö²æ¡¢ÆÀ²âϵͳ׷׏¤ÒµÐðʲ¹¿ÎµÄÔÓÂÒÆÚ¡£¡£¡£¡£¡£Ò²ÕýÓÉÓÚÔÓÂÒ£¬£¬£¬£¬£¬benchmark ²ÅÖµµÃ¿´£ºËüÃÇÊÇÕⳡ½ç˵ÕùÈ¡À£¬£¬£¬£¬ÉÙÊýÄÜ°Ñ¿ÚºÅÂäµ½²âÊÔÌâÉϵŤ¾ß¡£¡£¡£¡£¡£

Benchmark ¿ÉÒÔ¸øÒ»¸ö»¹ºÜËÉÉ¢µÄ¿´·¨»­½çÏß¡£¡£¡£¡£¡£WorldModelBench ¿¼²ìÊÓƵģ×ÓÓÐûÓÐ֪ʶºÍÎïÀíÔ¼Êø£»£»£»WorldScore Ȩºâ²î±ðÌìÏÂÌìÉúõ辶Äܲ»¿Éͳһ½ÏÁ¿£»£»£»WorldLens ¿ÉÒÔÓÃÀ´ÆÀ¹À¼ÝÊ»ÌìÏÂÊÇ·ñÕæµÄ¶ÔÇå¾²ºÍÏÂÓÎʹÃüÓÐÓ㻣»£»WorldArena ¿´µÄÊǾßÉíÌìÏÂÄ£×ÓÊÇ·ñÕæµÄÄÜ·þÎñ»úеÈË£»£»£»Physics-IQ Ôò°Ñ½çÏßѹµ½×îµ×²ã£¬£¬£¬£¬£¬ÎÊÄ£×Ó¾¿¾¹ÊÇÔÚÃ÷È·ÎïÀí£¬£¬£¬£¬£¬ÕÕ¾ÉÖ»ÊÇÔÚ²¹ÏñËØ¡£¡£¡£¡£¡£

ÌìÏÂÄ£×ÓÉú³¤µ½½ñÌ죬£¬£¬£¬£¬ÆÀ²âµÄÖ÷ÒªÐÔÕýÔÚÈÕÒæ͹ÏÔ£¬£¬£¬£¬£¬ÓïÑÔÄ£×ÓÒѾ­¸üÔç֤ʵÎúÕâ¼þÊ£¬£¬£¬£¬£¬½è¼øÓïÑÔÄ£×ÓµÄÂÄÀú£¬£¬£¬£¬£¬ÔçÈÕÈ·¶¨½¹µãÆÀ²âÆ«Ïò£¬£¬£¬£¬£¬Ò²½«³ÉΪδÀ´¾ºÕùµÄÒªº¦¡£¡£¡£¡£¡£

µã¸ö¡°°®ÐÄ¡±£¬£¬£¬£¬£¬ÔÙ×ß °É

ÆƽâÄÑÌâ·½Ãæ¡£¡£¡£¡£¡£Ëæן߱ÈÀýÐÂÄÜÔ´´ó¹æÄ£½ÓÈ룬£¬£¬£¬£¬µçÁ¦ÏµÍ³Ô½·¢Öش󣬣¬£¬£¬£¬Ðí¶à¹Å°åÒªÁìÄÑÒÔÓ¦¶ÔµÄÎÊÌ⣬£¬£¬£¬£¬È˹¤ÖÇÄÜ¿ÉÒԺܺýâ¾ö¡£¡£¡£¡£¡£ºÃ±È£¬£¬£¬£¬£¬µçÁ¦ÐÐÒµµÄ¡°Ô¦µç¡±´óÄ£×Ó£¬£¬£¬£¬£¬ÊµÊ±Ó¦¶Ô·çµç¡¢¹â·ü×ÅÁ¦µÄ²¨¶¯×ª±ä£¬£¬£¬£¬£¬¸¨ÖúÌìÉúµçÍøµ÷Àí·½°¸£¬£¬£¬£¬£¬¿ÉÒÔ½øÒ»²½ÌáÉýÐÂÄÜÔ´ÏûÄÉˮƽ¡£¡£¡£¡£¡£

¿­Ê±AG

³éѪ²éÖ×Áö±ê¼ÇÎï¿ÉÌæ»»³¦¾µÂð