AgentOptimizationCandidateEvaluation interface

Comparative evaluation summary for a candidate.

Properties

avg_latency_ms	Average end-to-end latency per task, rounded to milliseconds.
avg_tokens	Average total tokens consumed per task. Omitted when token usage was not measured.
completed_at	Timestamp when full candidate evaluation completed, represented in Unix time. Omitted when unavailable.
eval_id	Foundry evaluation identifier used to score this candidate. Omitted when unavailable.
eval_run_id	Foundry evaluation run identifier used to score this candidate. Omitted when unavailable.
score	Average composite score across evaluated tasks. Omitted until scoring produces an aggregate score.

Average end-to-end latency per task, rounded to milliseconds.

avg_latency_ms?: number

number

Average total tokens consumed per task. Omitted when token usage was not measured.

avg_tokens?: number

number

Timestamp when full candidate evaluation completed, represented in Unix time. Omitted when unavailable.

completed_at?: Date

Date

Foundry evaluation identifier used to score this candidate. Omitted when unavailable.

eval_id?: string

string

Foundry evaluation run identifier used to score this candidate. Omitted when unavailable.

eval_run_id?: string

string

Average composite score across evaluated tasks. Omitted until scoring produces an aggregate score.

score?: number

number