Benchmark
System One Mosaic Benchmark
A benchmark focused on evaluating the capabilities of agent decision engines. It includes multiple decision-oriented tasks and uses Borda count scoring to produce a comprehensive ranking of participating models.
Timeline (last 90 days)
Oct 12
Darwin-27B-ZTC-v2 ranked 1st out of 102 models on the System One Mosaic Benchmark decision engine leaderboard.
Unverified50%
Oct 12
The System One Mosaic Benchmark measures the suitability of models for agentic decision nodes, rather than general instruction-following or programming ability.
Unverified50%