Stats

Aggregate stats across every published reproduction. Cost-bearing tiles gate on records with measured spend; the rest count every published record.

99
Published
70 /99
Cost-bearing
$1.14
Median spend
$145.67
Total spent
Min
$0.23
P90
$5.89
Max
$11.11

Cost distribution

70 cost-bearing records bucketed by total spend.

<$0.10
0
$0.10–$0.50
20
$0.50–$1.00
12
$1.00–$2.00
17
$2.00–$5.00
13
≥$5.00
8

Spend by agent and model

Total dollars charged to each agent, split by LLM. Sum may exceed the cost-bearing total because per-record spend is fully attributed.

repro
$80.58
vuln_variant
$60.09
coding
$2.25
support
$2.11
judge
$0.65
accounts/fireworks/models/kimi-k2p5
accounts/fireworks/models/kimi-k2p6
claude-opus-4-7
gpt-5.1-codex
gpt-5.2-codex

By ecosystem

Count, median cost, and median duration per package ecosystem.

Ecosystem Count Median cost Median duration Share
npm 24 $0.43 12m 17s
pip 20 $1.58 15m 44s
composer 11 $0.55 13m 7s
github 11 $9.70 59m 16s
go 8 $1.26 23m 47s
c 5 $1.33 29m 29s
source 3 $1.68 16m 28s
Go module 1 38m 55s
Maven 1 48m 55s
PyPI 1 36m 8s
cargo 1 $4.18 33m 23s
cpp 1 47m 53s
gnu 1 35m 50s
pip (per GitHub advisory) 1 8m 19s
pypi 1 4m 50s
rubygems 1 $0.67 10m 15s
rust 1 $0.27 10m 6s

By severity

Count and median cost per CVSS severity bucket.

critical 29 · median $0.55
high 54 · median $1.31
medium 11 · median $2.24
low 1 · median $1.02