Run started 2026-10-17 09:36 UTC
· completed 09:43 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
106.78s
p50
93.56s
p90
131.63s
p99
142.03s
min
90.12s
max
143.19s
std
22.54s
First Insight Time
Time to the agent's first update, such as a tool call or a reasoning step.
mean
9.18s
p50
8.61s
p90
11.38s
p99
12.80s
min
6.82s
max
12.96s
std
2.27s
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
109.98s
p50
108.04s
p90
137.67s
p99
152.71s
min
68.57s
max
154.38s
std
24.68s
LLM Time
Total time inside LLM spans, summed across every LLM call in the case.
mean
73.31s
p50
71.26s
p90
85.12s
p99
85.70s
min
60.35s
max
85.76s
std
11.34s
Tool Time
Total time executing tools — database queries, API calls and the like.
mean
37.00s
p50
31.95s
p90
51.85s
p99
63.17s
min
24.71s
max
64.43s
std
15.67s
Other Time
Agent time outside the LLM and tools: orchestration overhead.
mean
0.11s
p50
0.00s
p90
0.30s
p99
0.35s
min
0.00s
max
0.36s
std
0.16s
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
67.7%
p50
68.3%
p90
73.3%
p99
73.9%
min
58.7%
max
74.0%
std
6.1%
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
33.0%
p50
32.5%
p90
40.3%
p99
43.7%
min
25.6%
max
44.1%
std
7.1%
Tool Calls
Number of tool spans the agent produced for the case.
mean
12.60
p50
12.00
p90
15.80
p99
16.88
min
9.00
max
17.00
std
3.05
sum
63.00
Total Tokens
LLM tokens consumed across every call in the case (prompt plus completion).