Models
Models under test
One card per model; scores come from its latest finished run. Simulated reference models are labelled — they exist so every view works without API keys.
Swift Budget (simulated)
sim:swift-budget
Omni
82.0
simulated2 runs
Capability
68
Reliability
72
Safety
90
Agency
94
Economics
86
last run 23s ago$0.00318 / run
Atlas Frontier (simulated)
sim:atlas-frontier
Omni
79.5
simulated1 run
Capability
96
Reliability
92
Safety
100
Agency
78
Economics
32
last run 24s ago$0.5756 / run
Local Llama 8B (simulated)
sim:local-llama
Omni
58.4
simulatedlocal2 runs
Capability
46
Reliability
62
Safety
40
Agency
53
Economics
90
last run 22s ago$0.00162 / run