Agentic x Wrong answer Ranking

See which AI models are most likely to hit Wrong answer on Agentic, so you can spot weak points faster. Sort by: Total Cost ↑.

Models Shown

Total Failures

108

Most Affected Model

Failure Reasons

Wrong answer108 Invalid tool call44 API error32 Did not follow instructions14 Timed out10

Categories

Domain specific766 Anti-AI Tricks446 Coding429 Puzzle Solving307 Trivia280 Combined115 General Intelligence112 Agentic108 Instructions following94 Data parsing and extraction71

108/108

Rank	Model	Company	Wrong answer Count	Category Score	Total Cost	Tests Correct	Response Time (avg)
#78	Claude Opus 4.8 medium	Anthropic	1	6.1	$3.490	0/1	99.4s
Total Tests 1 Wrong Tests 1 Total Cost $3.490 Response Time (avg) 99.4s
#134	Grok 4.7 medium	X AI	1	3.9	$3.501	0/1	372.1s
Total Tests 1 Wrong Tests 1 Total Cost $3.501 Response Time (avg) 372.1s
#142	Claude Opus 4.8 low	Anthropic	1	4.7	$3.787	0/1	110.7s
Total Tests 1 Wrong Tests 1 Total Cost $3.787 Response Time (avg) 110.7s

Filter models