Run started 2026-10-19 13:39 UTC
· completed 13:40 UTC
Headline Metrics
A case passes when (Correctness or String in Response) and Relevance and Tools pass.
Scorers that had nothing to judge are ignored; cases where no scorer produced a result are
counted as no data and excluded from every pass rate.
First Token Time
Time to the first response token (TTFT) — how long before the user sees anything.
mean
58.81s
p50
58.81s
p90
63.55s
p99
64.61s
min
52.89s
max
64.73s
std
8.37s
First Insight Time
Time to the agent's first update, such as a tool call or a reasoning step.
mean
7.34s
p50
7.34s
p90
8.45s
p99
8.70s
min
5.95s
max
8.73s
std
1.97s
Case Duration
Wall-clock time the eval harness measured for the whole case, including scoring and gRPC overhead. Always at least the agent's own time_total.
mean
8.32s
p50
0.02s
p90
30.56s
p99
69.92s
min
0.01s
max
71.50s
std
22.68s
LLM Time
Total time inside LLM spans, summed across every LLM call in the case.
mean
41.95s
p50
41.95s
p90
44.73s
p99
45.36s
min
38.46s
max
45.43s
std
4.93s
Tool Time
Total time executing tools — database queries, API calls and the like.
mean
18.31s
p50
18.31s
p90
18.72s
p99
18.81s
min
17.80s
max
18.82s
std
0.72s
Other Time
Agent time outside the LLM and tools: orchestration overhead.
mean
0.81s
p50
0.81s
p90
1.46s
p99
1.60s
min
0.00s
max
1.62s
std
1.15s
LLM Time Share
LLM time as a percentage of the agent's total span duration. This is the spread of per-case percentages, not the run-level ratio.
mean
69.8%
p50
69.8%
p90
70.4%
p99
70.6%
min
69.0%
max
70.6%
std
1.1%
Tool Time Share
Tool time as a percentage of the agent's total span duration. This is the spread of per-case percentages, not the run-level ratio.
mean
30.6%
p50
30.6%
p90
32.3%
p99
32.6%
min
28.6%
max
32.7%
std
2.9%
Tool Calls
Number of tool spans the agent produced for the case.
mean
10.50
p50
10.50
p90
10.90
p99
10.99
min
10.00
max
11.00
std
0.71
sum
21.00
Total Tokens
LLM tokens consumed across every call in the case (prompt plus completion).