CategoryWeight: 1.0x

Bug Fixes

Identify and fix bugs in existing codebases, including race conditions, off-by-one errors, and logic flaws.

Best Score

0.0

Avg Score

0.0

Tests

Performance Over Time — All Models

Model Rankings

Claude Opus 5

Category score

View

98.4BEST

Tokens13.9k

Total13.9k

Claude Sonnet 5

Category score

View

95.7-2.7 pts

Tokens6.2k

Total6.2k

Grok 4.5

Category score

View

92.7-5.7 pts

Tokens63.6k

Total63.6k

GPT-5.6 Sol

Category score

View

88.0-10.4 pts

Tokens45.1k

Total45.1k

Rank	Model	Score	Tokens	vs. Best	Details
1	Claude Opus 5	98.4	13.9k	BEST	View
2	Claude Sonnet 5	95.7	6.2k	-2.7 pts	View
3	Grok 4.5	92.7	63.6k	-5.7 pts	View
4	GPT-5.6 Sol	88.0	45.1k	-10.4 pts	View

Test Breakdown

Off-by-One Boundary Fix

Fix pagination logic that skips the last page of results

Claude Opus 5

95.2

Claude Sonnet 5

95.2

Grok 4.5

85.2

GPT-5.6 Sol

74.0

Race Condition Detection

Find and fix a subtle race condition in async queue processing

Claude Opus 5

100.0

Claude Sonnet 5

96.0

Grok 4.5

97.0

GPT-5.6 Sol

100.0

Memory Leak Fix

Identify and patch a memory leak caused by unclosed event listeners

Claude Opus 5

100.0

Claude Sonnet 5

96.0

Grok 4.5

96.0

GPT-5.6 Sol

90.0