Back to Archive

AutoBench Run 6 - August 2026

Twenty models scored on 252 questions across ten subject domains. Score, cost, and latency are the per-answer committee rank, billed cost, and generation time. Routers in the AutoBench Router Benchmark are scored against this answer matrix.

Latest
Date
August 29, 2026
Version
2026-08-29
Models
20
New Models
18

Run data

Model
Average (All Topics)Computer SciencePhilosophySocial ScienceLanguageScienceTechnologyArtsLiteratureHistoryAgentic
488.25s (#20)285.18s (#8)262.10s (#20)332.57s (#19)509.58s (#19)555.65s (#20)469.04s (#20)355.05s (#16)328.50s (#19)362.34s (#20)317.22s (#20)
487.51s (#19)514.94s (#18)185.84s (#18)210.49s (#15)559.35s (#20)268.77s (#10)378.15s (#18)361.90s (#17)167.32s (#16)153.11s (#18)171.78s (#19)
343.88s (#12)335.60s (#11)47.99s (#5)118.81s (#6)279.25s (#12)386.92s (#17)304.95s (#17)224.54s (#10)37.27s (#4)124.86s (#15)136.10s (#18)
322.09s (#11)322.53s (#9)148.41s (#14)130.90s (#10)354.53s (#14)255.64s (#8)453.09s (#19)273.67s (#13)60.63s (#6)45.38s (#5)97.26s (#17)
472.50s (#18)556.74s (#20)87.88s (#10)156.53s (#12)464.45s (#18)405.37s (#19)168.32s (#9)488.91s (#20)73.26s (#8)51.12s (#6)71.60s (#16)
298.19s (#10)350.95s (#12)160.19s (#17)305.56s (#18)244.04s (#11)232.04s (#7)259.32s (#13)252.52s (#11)212.75s (#17)142.50s (#17)70.10s (#15)
193.84s (#7)328.24s (#10)102.80s (#11)124.08s (#8)117.26s (#7)256.56s (#9)172.00s (#10)58.23s (#3)103.19s (#11)114.10s (#13)68.67s (#14)
400.97s (#16)427.59s (#15)158.64s (#16)446.85s (#20)369.60s (#16)364.33s (#15)270.72s (#16)277.75s (#14)340.12s (#20)160.89s (#19)58.05s (#13)
371.82s (#13)485.43s (#17)216.75s (#19)273.77s (#17)328.31s (#13)339.48s (#14)268.12s (#14)302.65s (#15)103.09s (#10)75.03s (#10)54.44s (#12)
237.80s (#8)516.02s (#19)63.22s (#6)173.83s (#14)167.89s (#8)394.91s (#18)124.42s (#7)216.79s (#9)103.35s (#12)86.53s (#12)44.64s (#11)
376.61s (#14)388.77s (#14)154.50s (#15)125.17s (#9)231.23s (#10)333.05s (#13)183.63s (#11)422.19s (#18)110.44s (#14)140.82s (#16)32.41s (#10)
408.55s (#17)256.71s (#7)147.18s (#13)158.09s (#13)404.50s (#17)300.07s (#12)269.86s (#15)452.31s (#19)156.35s (#15)123.57s (#14)26.68s (#9)
277.53s (#9)482.28s (#16)65.53s (#7)264.70s (#16)186.98s (#9)273.80s (#11)40.75s (#1)271.20s (#12)226.03s (#18)76.93s (#11)26.59s (#8)
112.41s (#2)90.68s (#2)72.44s (#9)56.28s (#3)85.18s (#5)127.40s (#4)102.84s (#5)70.58s (#4)33.56s (#3)42.28s (#4)25.84s (#7)
145.77s (#6)119.32s (#4)65.65s (#8)121.16s (#7)82.84s (#4)148.78s (#5)219.60s (#12)197.87s (#8)110.27s (#13)68.68s (#9)22.48s (#6)
381.49s (#15)381.29s (#13)104.85s (#12)150.73s (#11)360.23s (#15)382.06s (#16)140.66s (#8)95.99s (#5)92.07s (#9)60.93s (#8)21.86s (#5)
116.93s (#3)112.00s (#3)44.49s (#3)73.25s (#5)103.81s (#6)104.53s (#2)106.85s (#6)175.84s (#7)46.79s (#5)41.89s (#3)19.77s (#4)
142.98s (#5)139.08s (#5)21.79s (#2)32.97s (#2)77.17s (#3)190.17s (#6)63.61s (#4)34.02s (#1)32.05s (#2)21.19s (#1)18.08s (#3)
82.39s (#1)79.36s (#1)47.82s (#4)27.41s (#1)51.06s (#2)84.48s (#1)59.76s (#3)37.39s (#2)24.91s (#1)28.99s (#2)15.21s (#2)
117.31s (#4)150.80s (#6)19.39s (#1)72.74s (#4)38.92s (#1)117.35s (#3)43.92s (#2)128.34s (#6)71.17s (#7)54.03s (#7)11.05s (#1)
AutoBench Run 6 - August 2026 - AutoBench