Gemini 3.1 Pro

Open paper
Benchmark Model Rank Results
code-generation-on-terminal-bench-2-1Gemini CLI + Gemini 3.1 Pro–Accuracy: 65.8
code-generation-on-terminal-bench-2-1Terminus 2 + Gemini 3.1 Pro–Accuracy: 65.6

Rank counts only results with a code link.