Marketplace Agent Benchmark

VibeFuse · Widget

Agent Benchmark

Run MMLU-Pro, GSM8K, HumanEval, and a VibeFuse agent exam against the selected model. Knowledge suites are tools-off and comparable to published accuracy %. HumanEval and the agent exam allow terminal, widgets, and tools. Returns per-suite percentages plus a headline VibeFuse Agent Score.

Sign in and save to your library — VibeFuse syncs free widgets into the Windows canvas on next launch.

v1.2.0 custom 0 saves 1 views Free
Sign in to save Get VibeFuse

About this widget

Run MMLU-Pro, GSM8K, HumanEval, and a VibeFuse agent exam against the selected model. Knowledge suites are tools-off and comparable to published accuracy %. HumanEval and the agent exam allow terminal, widgets, and tools. Returns per-suite percentages plus a headline VibeFuse Agent Score.