Anti-AI Tricks x Wrong answer Ranking

AI BENCHY Category Failures

See which AI models are most likely to hit Wrong answer on Anti-AI Tricks, so you can spot weak points faster. Sort by: Response Time (avg) ↑.

Models Shown

Total Failures

245

Most Affected Model

Failure Reasons

Wrong answer245 Did not follow instructions30 Extra formatting14 API error13 Timed out4 No answer2

Categories

Domain specific314 Anti-AI Tricks245 Coding194 Puzzle Solving147 Trivia130 Instructions following53 Combined52 Data parsing and extraction35 General Intelligence32 Tool Calling2

Rank	Model	Company	Wrong answer Count	Category Score	Tests Correct	Response Time (avg)
#90	Gemini 3.1 Flash Lite none	Google	2	7.5	2/4	1.07s
#157	Grok 4.1 Fast none	X AI	3	3.2	0/4	1.07s
#34	Qwen3.7 Max none	Qwen	2	6.5	2/4	1.08s
#87	Gemini 3.1 Flash Lite minimal	Google	1	8.3	3/4	1.10s
#81	Mercury 2 medium	Inception	1	6.9	2/4	1.12s
#148	GPT-5.4 Nano none	OpenAI	4	3.5	0/4	1.18s
#136	Elephant Alpha medium	Openrouter	2	6.6	2/4	1.19s
#152	MiMo-V2-Flash none	Xiaomi	4	3.2	0/4	1.19s
#149	Nemotron 3 Nano Omni 30b A3b Reasoning medium	NVIDIA	1	6.4	2/4	1.20s
#125	GPT-5.4 none	OpenAI	4	3.2	0/4	1.21s
#48	Gemini 3 Flash Preview none	Google	1	8.3	3/4	1.25s
#102	Gemma 4 26B A4B none	Google	1	8.3	3/4	1.28s
#91	GPT-5.5 none	OpenAI	2	6.9	2/4	1.31s
#108	Qwen3.5-Flash none	Qwen	4	3.5	0/4	1.32s
#147	GPT-4o-mini none	OpenAI	3	4.8	1/4	1.34s

←

1 2 3 8

→

Top Models by Wrong answer Count