Live Benchmarks

LanceDB Benchmark

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

View on GitHubTotal tasks: 10Last run: 7/23/2026

Model Performance

ModelPassedAvg DurationSuccess Rate
#1
gemini-3.5-flashNEW
9285.4s
90%
#2
claude-4-6-sonnet
8162.4s
80%
#3
glm-5.1
7287.1s
70%
#4
gemini-3-flash
7129.6s
70%
#5
gemini-3.1-pro
6153.4s
60%