BridgeBenchBridgeBench
Refactoring
Model Analysis

DeepSeek V4 Pro

openrouter/deepseek/deepseek-v4-pro

48.7

overall score

73.3% visible
72.7% hidden
61.9% intent

Tasks

15

Passed

10

Failed

5

Avg latency

43388ms

Total cost

$0.1019

Cluster Performance

Control Flow3 tasks
79.4
Data Pipelines3 tasks
50.9
State Isolation3 tasks
25.4
Duplication3 tasks
35.4
Modernization3 tasks
52.5

All Task Results

TaskClusterScore
Transaction Aggregation Without Mutable Accumulators

refactor-global-state

State Isolation0.0
Named Conversion Constants

refactor-magic-numbers

Modernization0.0
Declarative Validation Rules

refactor-validation-rules

Duplication0.0
ETL Record Transformation

refactor-data-transformer

Data Pipelines17.6
Event Handler Registry

refactor-event-handler

State Isolation35.3
Inventory Operations by Handler

refactor-class-to-functions

State Isolation41.0
Composable Sales Report Generation

refactor-report-generator

Duplication43.5
Guard Clauses for Nested Categorization

refactor-nested-conditionals

Control Flow54.8
Order Processing Orchestration

refactor-god-function

Modernization62.0
Shared Quarter Ranking Logic

refactor-duplicate-code

Duplication62.7
Composable Department Pipeline

refactor-promise-chain

Data Pipelines62.7
Student Classification Pipeline

refactor-array-manipulation

Data Pipelines72.3
Operation Map Calculator

refactor-calculator

Control Flow86.7
Template Literal Contact Card

refactor-string-concat

Modernization95.5
Lookup Table for Status Text

refactor-switch-to-map

Control Flow96.6

15tasks · Sorted by score (lowest first) · Intent = structural refactor compliance