Back to leaderboard

google-adk

PythonLlmAgent (LiteLlm wrapper)
Model: gemini-2.5-flashGenerated: May 11, 2026, 12:00 AM UTC
NDCG@3
0.621
29 scored · 29/30 valid
Hit@1
72.4%
JustifQ 3.41/5
Latency p50
19.9s
p95 471.8s
Mean tokens
6,638
6,128 in · 510 out
Cost / run
$0.0184
9.3 avg tool calls

Discussed in /findings

All metrics

count_total
30
count_valid
29
success_rate
96.7%
latency_p50
19.858s
latency_p95
471.806s
latency_mean
93.505s
latency_max
752.416s
mean_input_tokens
6,128
mean_output_tokens
510
mean_tool_calls
9.34
estimated_cost_usd_per_run
$0.018387
mean_ndcg_at_3
0.621
hit_at_1_rate
72.4%
mean_precision_at_3
0.356
mean_recall_at_3
0.460
n_scored
29
mean_justification_quality
3.41/5
mean_judge_score
15.03/20
judge_n
29
hit_step_limit_rate
0.0%