Run started 2026-10-18 07:32 UTC
· completed 07:38 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
74.42s
p50
70.62s
p90
105.99s
p99
120.58s
min
42.04s
max
122.20s
std
26.94s
First Insight Time
Time to the agent's first update, such as a tool call or a reasoning step.
mean
10.04s
p50
10.24s
p90
11.38s
p99
13.27s
min
7.08s
max
13.48s
std
1.79s
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
69.22s
p50
71.26s
p90
133.43s
p99
163.53s
min
6.42s
max
168.33s
std
48.64s
LLM Time
Total time inside LLM spans, summed across every LLM call in the case.
mean
47.76s
p50
45.95s
p90
63.07s
p99
91.92s
min
29.64s
max
95.13s
std
20.03s
Tool Time
Total time executing tools — database queries, API calls and the like.
mean
29.43s
p50
25.37s
p90
55.24s
p99
67.43s
min
7.67s
max
68.78s
std
19.37s
Other Time
Agent time outside the LLM and tools: orchestration overhead.
mean
2.45s
p50
2.57s
p90
5.55s
p99
5.68s
min
0.00s
max
5.69s
std
2.19s
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
62.4%
p50
64.0%
p90
73.5%
p99
76.3%
min
30.5%
max
76.6%
std
14.6%
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
35.5%
p50
32.3%
p90
53.9%
p99
69.2%
min
17.7%
max
70.9%
std
16.4%
Tool Calls
Number of tool spans the agent produced for the case.
mean
10.33
p50
11.00
p90
13.60
p99
15.76
min
7.00
max
16.00
std
3.08
sum
93.00
Total Tokens
LLM tokens consumed across every call in the case (prompt plus completion).