Did not follow instructions Failure Ranking

See which AI models run into Did not follow instructions most often, so you can spot reliability risks before choosing one. Sort by: Score ↑.

Models Shown

Total Failures

386

Most Affected Model

Categories

In category Puzzle Solving128 In category General Intelligence115 In category Instructions following48 In category Anti-AI Tricks44 In category Coding29 In category Tool Calling12 In category Data parsing and extraction7 In category Domain specific2 In category Combined1

215/215

Rank	Model	Company	Did not follow instructions Count	Score	Total Cost	Tests Correct	Response Time (avg)
#32	Seed 2.1 Turbo high	Bytedance Seed	1	8.9	$1.797	16/22	174.0s
Total Tests 22 Wrong Tests 6 Total Cost $1.797 Response Time (avg) 174.0s
#31	Muse Spark 1.3 high	Meta	2	8.9	$1.653	16/22	52.5s
Total Tests 22 Wrong Tests 6 Total Cost $1.653 Response Time (avg) 52.5s
#30	GPT-5.3-Codex medium	OpenAI	2	8.9	$0.988	16/22	16.9s
Total Tests 22 Wrong Tests 6 Total Cost $0.988 Response Time (avg) 16.9s
#29	Grok 4.5 high	X AI	1	9.0	$2.252	17/22	72.3s
Total Tests 22 Wrong Tests 5 Total Cost $2.252 Response Time (avg) 72.3s
#26	Gemini 3.5 Flash medium	Google	1	9.0	$0.665	18/22	8.48s
Total Tests 22 Wrong Tests 4 Total Cost $0.665 Response Time (avg) 8.48s

Did not follow instructions Failures