Live Benchmarks

Amika Benchmark

Performance results of AI coding models on Amika tasks, measuring success rate and execution time with high precision.

View on GitHubTotal tasks: 33Last run: 4/27/2026

Model Performance

ModelPassedAvg DurationSuccess Rate
#1
gemini-3.1-proNEW
4531.4s
29%
#2
gemini-3-flash
8488.1s
24%