Run started 2026-10-17 08:54 UTC
· completed 09:00 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
70.59s
p50
58.83s
p90
111.21s
p99
132.45s
min
31.15s
max
135.41s
std
30.68s
First Insight Time
Time to the agent's first update, such as a tool call or a reasoning step.
mean
7.52s
p50
7.75s
p90
10.62s
p99
11.43s
min
3.24s
max
11.55s
std
2.53s
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
84.22s
p50
66.86s
p90
135.87s
p99
155.43s
min
45.96s
max
156.19s
std
34.30s
LLM Time
Total time inside LLM spans, summed across every LLM call in the case.
mean
61.21s
p50
51.68s
p90
100.28s
p99
120.84s
min
29.36s
max
122.73s
std
27.97s
Tool Time
Total time executing tools — database queries, API calls and the like.
mean
13.49s
p50
12.32s
p90
24.08s
p99
29.53s
min
3.68s
max
30.09s
std
7.97s
Other Time
Agent time outside the LLM and tools: orchestration overhead.
mean
0.75s
p50
0.00s
p90
2.16s
p99
4.43s
min
0.00s
max
4.79s
std
1.39s
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
83.0%
p50
83.8%
p90
93.5%
p99
99.4%
min
64.7%
max
100.1%
std
9.8%
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
18.4%
p50
18.3%
p90
29.6%
p99
36.0%
min
8.4%
max
36.5%
std
8.5%
Tool Calls
Number of tool spans the agent produced for the case.
mean
10.20
p50
9.00
p90
14.20
p99
15.00
min
6.00
max
15.00
std
3.08
sum
153.00
Total Tokens
LLM tokens consumed across every call in the case (prompt plus completion).