Savings and performance
Baseline cost is an estimate: the same tokens priced on the reference model for that task. Only routed requests are included.
Estimated savings
—
Success rate
Routed latency
First token —
Throughput
tokens/sec · 0 lag switches
Measured token savings
These amounts are charges Optimal API did not take. Routing savings above stay an estimate against the baseline model.
Total measured: Not measured yet
Response cache: Not measured yet
Prompt cache discount: Not measured yet
Context trimmed: Not measured yet
Prompt compression: Not measured yet
Batch discount: Not measured yet
Volume or commit discount: Not measured yet
Max tokens lowered: Not measured yet
Estimated savings by day
Nothing to chart in this window.
Latency comparison
Routed average latency: —
Routed average tokens/sec: —
Baseline model average latency: Not measured yet. This appears when the same account also calls the baseline model directly.
Cost by task
Nothing to chart in this window.
No routed tasks yet.
Models by task
No routed models yet.
Lag switches
No lag switches in this window.