Back to Archive

AutoBench Agentic Run 1 - April 2026

The first AutoBench run to measure agentic performance of top LLMs

Latest
Date
April 19, 2026
Version
2026-04-19
Models
32
New Models
1

Run data

Model
Average (All Topics)Adaptive ReplanningApi WorkflowDomain WorkflowError HandlingFailure RecoveryMulti Step OrchestrationParallel ExecutionParameter ComplexitySingle Tool CallTool Selection
21.55s (#1)26.34s (#3)27.95s (#3)30.02s (#3)21.89s (#2)21.69s (#1)15.67s (#1)29.06s (#3)18.66s (#4)10.02s (#1)14.23s (#2)
22.53s (#2)25.08s (#2)27.29s (#2)25.37s (#1)23.12s (#3)26.88s (#2)19.43s (#2)24.81s (#1)22.61s (#6)18.78s (#5)11.92s (#1)
30.73s (#3)34.73s (#5)51.20s (#8)34.91s (#5)37.86s (#5)30.83s (#3)29.59s (#4)43.74s (#4)13.65s (#2)14.12s (#3)16.67s (#5)
36.45s (#4)40.97s (#7)34.92s (#5)36.33s (#6)30.20s (#4)46.45s (#4)32.10s (#5)78.24s (#9)17.63s (#3)25.20s (#8)22.47s (#7)
41.00s (#5)28.19s (#4)38.43s (#6)33.85s (#4)46.89s (#6)47.65s (#5)47.89s (#8)106.97s (#14)22.32s (#5)22.47s (#6)15.30s (#4)
41.25s (#6)22.59s (#1)18.28s (#1)26.93s (#2)17.24s (#1)243.06s (#27)20.71s (#3)26.23s (#2)8.45s (#1)11.09s (#2)17.96s (#6)
47.01s (#7)37.49s (#6)63.41s (#11)58.66s (#9)64.33s (#11)47.68s (#6)33.14s (#6)86.67s (#10)33.18s (#7)17.99s (#4)27.53s (#8)
54.67s (#8)45.97s (#9)69.68s (#12)43.21s (#7)53.61s (#7)57.12s (#7)64.15s (#11)72.84s (#7)49.35s (#12)61.98s (#17)28.82s (#9)
57.54s (#9)70.46s (#13)63.29s (#10)70.59s (#12)73.86s (#13)64.28s (#8)60.17s (#10)62.53s (#5)45.93s (#10)32.23s (#9)32.07s (#10)
58.21s (#10)44.94s (#8)53.45s (#9)43.39s (#8)55.13s (#9)86.87s (#14)51.50s (#9)113.27s (#15)46.15s (#11)43.25s (#11)44.14s (#12)
63.36s (#11)51.60s (#10)33.31s (#4)73.13s (#13)66.80s (#12)76.08s (#11)43.41s (#7)91.52s (#12)56.49s (#14)126.51s (#24)14.77s (#3)
76.10s (#12)64.54s (#12)73.80s (#13)67.02s (#11)63.99s (#10)82.16s (#12)118.40s (#18)76.50s (#8)72.38s (#17)68.62s (#20)73.61s (#21)
78.57s (#13)60.92s (#11)180.76s (#20)61.16s (#10)111.80s (#16)70.48s (#9)86.79s (#12)72.18s (#6)60.30s (#15)40.20s (#10)41.12s (#11)
95.82s (#14)98.08s (#17)86.55s (#14)110.61s (#18)54.03s (#8)88.82s (#15)236.05s (#26)86.72s (#11)73.69s (#18)56.25s (#15)67.38s (#19)
97.92s (#15)84.54s (#15)197.75s (#26)100.16s (#17)80.42s (#14)86.53s (#13)111.44s (#17)150.53s (#19)70.63s (#16)44.96s (#14)52.25s (#14)
106.88s (#16)99.61s (#18)113.88s (#18)180.36s (#22)163.17s (#19)75.95s (#10)90.04s (#14)114.39s (#16)43.05s (#8)90.51s (#23)97.86s (#24)
114.39s (#17)158.95s (#21)43.21s (#7)83.20s (#15)191.29s (#22)237.75s (#26)119.98s (#19)140.97s (#17)86.19s (#19)24.62s (#7)57.78s (#17)
129.25s (#18)188.98s (#25)112.56s (#17)79.21s (#14)88.07s (#15)259.09s (#29)250.29s (#28)152.20s (#20)49.59s (#13)56.83s (#16)55.67s (#16)
134.63s (#19)129.62s (#19)110.14s (#15)95.46s (#16)303.95s (#29)131.78s (#17)89.62s (#13)212.00s (#24)160.59s (#27)63.75s (#18)49.37s (#13)
138.83s (#20)174.36s (#23)185.77s (#22)131.29s (#20)142.23s (#17)201.72s (#22)135.57s (#20)185.40s (#22)118.55s (#23)43.37s (#12)70.04s (#20)
149.12s (#21)154.34s (#20)191.42s (#24)279.76s (#30)153.42s (#18)208.83s (#24)106.87s (#16)141.66s (#18)147.49s (#26)43.59s (#13)63.82s (#18)
151.05s (#22)80.59s (#14)110.27s (#16)127.08s (#19)276.19s (#27)113.23s (#16)241.89s (#27)95.15s (#13)97.59s (#20)78.06s (#22)290.47s (#30)
165.52s (#23)185.96s (#24)182.14s (#21)169.19s (#21)208.18s (#24)174.71s (#19)141.33s (#21)205.74s (#23)137.31s (#24)145.45s (#25)105.24s (#25)
174.19s (#24)94.18s (#16)192.00s (#25)250.36s (#26)188.65s (#21)168.30s (#18)141.91s (#22)283.42s (#29)43.56s (#9)65.94s (#19)313.53s (#31)
177.55s (#25)334.16s (#30)122.39s (#19)189.85s (#23)167.28s (#20)191.31s (#20)97.54s (#15)274.55s (#26)139.53s (#25)177.26s (#26)81.63s (#22)
183.27s (#26)167.48s (#22)251.43s (#30)275.43s (#29)192.38s (#23)233.09s (#25)168.08s (#23)281.39s (#27)104.81s (#21)72.48s (#21)86.12s (#23)
240.81s (#27)240.50s (#26)217.96s (#28)268.45s (#27)395.57s (#32)245.20s (#28)268.93s (#29)238.99s (#25)112.89s (#22)200.63s (#27)219.00s (#28)
240.88s (#28)333.09s (#29)217.50s (#27)303.03s (#32)239.12s (#25)302.55s (#32)198.36s (#24)179.33s (#21)267.11s (#30)313.20s (#30)55.52s (#15)
245.21s (#29)346.35s (#31)232.45s (#29)206.61s (#25)305.13s (#30)295.02s (#31)227.88s (#25)315.69s (#30)169.17s (#28)212.15s (#28)141.62s (#26)
262.21s (#30)254.31s (#27)188.87s (#23)196.32s (#24)291.05s (#28)196.74s (#21)301.91s (#30)327.64s (#32)320.11s (#31)320.85s (#31)224.27s (#29)
284.56s (#31)312.20s (#28)357.31s (#32)290.42s (#31)252.05s (#26)278.36s (#30)387.20s (#32)282.72s (#28)200.52s (#29)267.32s (#29)217.48s (#27)
325.77s (#32)356.84s (#32)304.61s (#31)268.49s (#28)393.60s (#31)205.17s (#23)323.54s (#31)321.91s (#31)396.14s (#32)334.06s (#32)353.36s (#32)
AutoBench Agentic Run 1 - April 2026 - AutoBench