12 comments
12 forecasters
Which of these AI benchmarks will first have its record beaten by 10%?
SWE-bench16.7%
FrontierMathresult: Yes
LastExam28.5%
None, before July 4th23%
Forecast Timeline
Peer Score
community
+1.4
Average
of peers
Baseline Score
community
+10.2
0
Chance
baseline
Author:
Opened:Apr 25, 2026
Closes:Apr 25, 2026
Resolved:Jul 10, 2026
Spot Scoring Time:Apr 25, 2026
Include Bots:Yes