Run started 2026-10-19 04:09 UTC
· completed 04:10 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
19.95s
p50
15.12s
p90
27.18s
p99
29.90s
min
14.54s
max
30.20s
std
8.88s
First Insight Time
Time to the agent's first update, such as a tool call or a reasoning step.
mean
8.21s
p50
7.31s
p90
9.61s
p99
10.13s
min
7.12s
max
10.19s
std
1.72s
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
28.91s
p50
27.16s
p90
35.20s
p99
37.01s
min
22.34s
max
37.22s
std
7.59s
LLM Time
Total time inside LLM spans, summed across every LLM call in the case.
mean
20.00s
p50
20.52s
p90
29.02s
p99
30.93s
min
8.35s
max
31.14s
std
11.40s
Tool Time
Total time executing tools — database queries, API calls and the like.
mean
2.32s
p50
1.04s
p90
4.78s
p99
5.63s
min
0.20s
max
5.72s
std
2.97s
Other Time
Agent time outside the LLM and tools: orchestration overhead.
mean
0.47s
p50
0.00s
p90
1.13s
p99
1.38s
min
0.00s
max
1.41s
std
0.81s
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
88.2%
p50
99.1%
p90
109.0%
p99
111.3%
min
53.9%
max
111.5%
std
30.3%
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
13.8%
p50
3.3%
p90
30.2%
p99
36.3%
min
1.1%
max
37.0%
std
20.1%
Tool Calls
Number of tool spans the agent produced for the case.
mean
5.67
p50
4.00
p90
8.80
p99
9.88
min
3.00
max
10.00
std
3.79
sum
17.00
Total Tokens
LLM tokens consumed across every call in the case (prompt plus completion).