Adaptive Router
Thompson-sampling contextual bandit; feedback updates per-tenant Beta posteriors online.
#1claude-sonnet
α1/β150.0%
#2gemini-flash
α1/β150.0%
#3opus-thinking
α1/β150.0%
#4gpt-5.5
α1/β150.0%
Per-tenant policy
Rewards observed: 0
Persisted to
aios:router:policyRecent feedback
No feedback yet.