Run started 2026-10-17 09:52 UTC
· completed 10:00 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
103.60s
p50
103.33s
p90
131.86s
p99
147.59s
min
53.49s
max
149.34s
std
27.81s
First Insight Time
Time to the agent's first update, such as a tool call or a reasoning step.
mean
8.01s
p50
7.19s
p90
12.07s
p99
14.02s
min
3.33s
max
14.24s
std
3.83s
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
119.53s
p50
129.23s
p90
153.71s
p99
155.14s
min
62.00s
max
155.29s
std
30.05s
LLM Time
Total time inside LLM spans, summed across every LLM call in the case.
mean
81.08s
p50
85.98s
p90
97.13s
p99
130.19s
min
37.02s
max
133.86s
std
24.61s
Tool Time
Total time executing tools — database queries, API calls and the like.
mean
28.81s
p50
29.16s
p90
37.89s
p99
41.53s
min
17.17s
max
41.93s
std
8.65s
Other Time
Agent time outside the LLM and tools: orchestration overhead.
mean
0.27s
p50
0.00s
p90
0.00s
p99
2.71s
min
0.00s
max
3.01s
std
0.91s
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
74.8%
p50
72.2%
p90
84.9%
p99
88.6%
min
66.0%
max
89.0%
std
7.1%
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
27.9%
p50
30.7%
p90
32.9%
p99
39.7%
min
11.4%
max
40.4%
std
7.9%
Tool Calls
Number of tool spans the agent produced for the case.
mean
13.55
p50
14.00
p90
16.00
p99
16.00
min
11.00
max
16.00
std
2.02
sum
149.00
Total Tokens
LLM tokens consumed across every call in the case (prompt plus completion).