Run started 2026-10-19 11:08 UTC
· completed 13:14 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
56.25s
p50
50.37s
p90
79.28s
p99
124.03s
min
22.85s
max
124.81s
std
24.07s
First Insight Time
Time to the agent's first update, such as a tool call or a reasoning step.
mean
8.28s
p50
7.72s
p90
14.28s
p99
17.15s
min
3.48s
max
17.82s
std
4.08s
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
50.64s
p50
0.34s
p90
153.75s
p99
262.50s
min
0.02s
max
305.69s
std
74.21s
LLM Time
Total time inside LLM spans, summed across every LLM call in the case.
mean
47.87s
p50
41.62s
p90
64.81s
p99
111.08s
min
17.93s
max
112.92s
std
22.57s
Tool Time
Total time executing tools — database queries, API calls and the like.
mean
9.97s
p50
8.90s
p90
16.28s
p99
25.53s
min
2.72s
max
27.46s
std
5.48s
Other Time
Agent time outside the LLM and tools: orchestration overhead.
mean
1.51s
p50
1.18s
p90
4.01s
p99
6.13s
min
0.00s
max
6.25s
std
1.74s
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
81.4%
p50
82.7%
p90
94.4%
p99
101.6%
min
51.9%
max
101.6%
std
11.2%
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
17.7%
p50
17.3%
p90
26.4%
p99
44.8%
min
7.7%
max
48.3%
std
8.9%
Tool Calls
Number of tool spans the agent produced for the case.
mean
10.86
p50
11.00
p90
13.00
p99
16.00
min
7.00
max
16.00
std
2.13
sum
315.00
Total Tokens
LLM tokens consumed across every call in the case (prompt plus completion).