CategoryWeight: 1.0x

Long Reasoning

Multi-step logic puzzles, extended chain-of-thought, and complex analytical reasoning tasks requiring sustained coherence over many steps.

Best Score

0.0

Avg Score

0.0

Tests

Performance Over Time — All Models

Model Rankings

Claude Opus 4.8

Category score

View

77.8BEST

Tokens10.7k

Total10.7k

Grok

Category score

View

70.1-7.7 pts

Tokens92.2k

Total92.2k

Claude Sonnet 4.6

Category score

View

68.9-8.9 pts

Tokens20.0k

Total20.0k

GPT-5.5

Category score

View

66.6-11.2 pts

Tokens38.1k

Total38.1k

Rank	Model	Score	Tokens	vs. Best	Details
1	Claude Opus 4.8	77.8	10.7k	BEST	View
2	Grok	70.1	92.2k	-7.7 pts	View
3	Claude Sonnet 4.6	68.9	20.0k	-8.9 pts	View
4	GPT-5.5	66.6	38.1k	-11.2 pts	View

Test Breakdown

Multi-step Logic Puzzle

Complex optimization with 8+ constraints across multiple variables

Claude Opus 4.8

77.8

Grok

70.1

Claude Sonnet 4.6

68.9

GPT-5.5

66.6

Legal Reasoning Chain

Contract dispute analysis requiring multi-party obligation tracking

Claude Opus 4.8

77.8

Grok

70.1

Claude Sonnet 4.6

68.9

GPT-5.5

66.6

Mathematical Proof

Prove divisibility properties using induction and modular arithmetic

Claude Opus 4.8

77.8

Grok

70.1

Claude Sonnet 4.6

68.9

GPT-5.5

66.6