BLACKBOX
flight recorder for AI agents

Performance Leaderboard

LIVE · FROM PRODUCTION TRAFFIC
Quality score, fault rate, and cost efficiency ranked from your actual recorded runs — not benchmarks. Every number is sourced from the immutable event log.
COST FORECAST — THIS MONTH
$0.3509
MONTH-TO-DATE
$0.0195
DAILY BURN RATE
$0.5848
PROJECTED MONTH END
18 / 30
DAYS ELAPSED
MODELMTDPROJECTED
claude-sonnet-4-6$0.32271$0.53785
claude-haiku-4-5-20251001$0.02817$0.04695
ACTIVITY HEATMAP — LAST 91 DAYS37 total runs
Sun
Mon
Tue
Wed
Thu
Fri
Sat
lessmore
TOTAL RUNS
37
TOTAL EVENTS
311
TOTAL COST
$0.3509
OVERALL FAULT RATE
42.9%
MODELS TRACKED
2
◈ CURRENT CHAMPION · RANKED #1 FROM 10 RUNS
claude-haiku-4-5-20251001
17 events recorded · 9 verified outputs
100.0
/ 100
★ new
FAULT RATE
0.0%(0/9)
COST / VERIFIED OUTPUT
$0.003130
AVG LATENCY
5,874ms
TOTAL COST
$0.0282
AVG TOKENS IN→OUT
1,542→587
MODEL RANKINGS — QUALITY SCORE · FAULT RATE · COST EFFICIENCY
RANKMODELQUALITY SCORETRENDFAULT RATERUNSCOST/RUNCOST/VERIFIEDAVG LATENCYTOKENS IN→OUTFAULT TRENDCHALLENGE
1
claude-haiku-4-5-20251001
CHAMPION
100.0
★ new0.0%(0/9)10$0.0028$0.0031305,874ms1,542→587vs →
2
claude-sonnet-4-6
73.3
★ new26.7%(4/15)21$0.0154$0.0215145,960ms1,661→490vs →
DAILY ACTIVITY — LAST 5 DAYS■ clean■ had faults
RUNS / DAY
COST / DAY
FAULTS / DAY
EVENTS / DAY
2026-06-17
3 runs · 9 events · $0.0000 · 2 faults
2026-06-16
21 runs · 227 events · $0.2415 · 3 faults
2026-06-14
8 runs · 32 events · $0.0000 · 6 faults
2026-06-13
4 runs · 41 events · $0.1094 · 4 faults
2023-11-14
1 run · 2 events · $0.0000
TOP RUNS BY COST — HALL OF RECORD
RUN IDMODEL SNAPSHOTEVENTSCOSTFAULTSREPLAY
run_meridian_003claude-sonnet-4-6@sn-4.6-051920$0.05723replay →
code_review_91b62f7026$0.03760replay →
run_meridian_001claude-sonnet-4-6@sn-4.6-051910$0.02861replay →
esg_carbon_eacf768f15$0.02230replay →
agent_run_fda39d629912$0.02180replay →
credit_risk_78d73eb512$0.02060replay →
fraud_monitor_e94580c410$0.02030replay →
compliance_gdpr_57b3702410$0.01950replay →
model_cost_3607ffb617$0.01940replay →
supply_chain_94d8774411$0.01930replay →
QUALITY SCORE vs AVG COST PER RUN
100%0%$0$0.02151avg cost per verified output →quality score ↑claude-haiku-4-5-2claude-sonnet-4-6
computed Thu, 18 Jun 2026 03:02:02 GMT