Instructions following Model Ranking

See which AI models perform best on Instructions following, which ones stay reliable, and where the biggest gaps appear. Sort by: Response Time (avg) ↑.

Models Shown

Average Instructions following Score

8.5

Best Model

Granite 4.1 8B 3.6

Failure Reasons

With failure reason Wrong answer82 With failure reason Did not follow instructions48 With failure reason Extra formatting7 With failure reason Timed out5 With failure reason API error2 With failure reason No answer2

316/316

Rank	Model	Company	Instructions following Score	Score	Total Cost	Tests Correct	Response Time (avg)
#102	GPT-5.4 Mini medium	OpenAI	9.8	7.5	$0.786	2/2	2.13s
Total Tests 2 Wrong Tests 0 Total Cost $0.786 Response Time (avg) 2.13s
#262	GLM 5 Turbo none	Z.ai	6.5	5.1	$0.047	1/2	2.13s
Total Tests 2 Wrong Tests 1 Total Cost $0.047 Response Time (avg) 2.13s
#70	GPT-5.6 Terra high	OpenAI	10.0	8.0	$0.836	2/2	2.14s
Total Tests 2 Wrong Tests 0 Total Cost $0.836 Response Time (avg) 2.14s
#35	Gemini 3.5 Flash Lite high	Google	9.8	8.8	$0.540	2/2	2.16s
Total Tests 2 Wrong Tests 0 Total Cost $0.540 Response Time (avg) 2.16s
#8	GPT-5.6 Sol low	OpenAI	10.0	9.5	$0.357	2/2	2.27s
Total Tests 2 Wrong Tests 0 Total Cost $0.357 Response Time (avg) 2.27s
#108	GPT-5.6 Luna medium	OpenAI	9.9	7.4	$0.072	2/2	2.38s
Total Tests 2 Wrong Tests 0 Total Cost $0.072 Response Time (avg) 2.38s
#119	DeepSeek V4 Flash 0731 medium	DeepSeek	8.5	7.3	$0.066	1/2	2.42s
Total Tests 2 Wrong Tests 1 Total Cost $0.066 Response Time (avg) 2.42s
#74	Qwen3.8 27B low	Qwen	10.0	7.9	~$0.071	2/2	2.43s
Total Tests 2 Wrong Tests 0 Total Cost ~$0.071 Response Time (avg) 2.43s
#87	Claude Opus 4.6 medium	Anthropic	10.0	7.7	$2.961	2/2	2.43s
Total Tests 2 Wrong Tests 0 Total Cost $2.961 Response Time (avg) 2.43s
#302	Ling 3.0 Tiny none	Inclusionai	3.1	4.0	$0.000	0/2	2.44s
Total Tests 2 Wrong Tests 2 Total Cost $0.000 Response Time (avg) 2.44s
#19	Gemini 3.7 Flash low	Google	10.0	9.2	$0.207	2/2	2.47s
Total Tests 2 Wrong Tests 0 Total Cost $0.207 Response Time (avg) 2.47s
#13	GPT-5.6 Sol medium	OpenAI	10.0	9.4	$0.508	2/2	2.50s
Total Tests 2 Wrong Tests 0 Total Cost $0.508 Response Time (avg) 2.50s
#229	Mimo V2 PRO none	Xiaomi	6.5	5.6	$0.045	1/2	2.51s
Total Tests 2 Wrong Tests 1 Total Cost $0.045 Response Time (avg) 2.51s
#115	KAT-Coder-Pro V2.5 low	Kwaipilot	10.0	7.3	$0.400	2/2	2.53s
Total Tests 2 Wrong Tests 0 Total Cost $0.400 Response Time (avg) 2.53s
#81	Qwen3.8 27B medium	Qwen	10.0	7.8	~$0.058	2/2	2.54s
Total Tests 2 Wrong Tests 0 Total Cost ~$0.058 Response Time (avg) 2.54s

Instructions following Ranking

Filter models

Top Models by Instructions following Score

Instructions following Score vs Total Cost

Top Models by Response Time (avg)

Instructions following Ranking

Filter models

Top Models by Instructions following Score

Instructions following Score vs Total CostEstimated GPU electricity cost only. It excludes the rest of the computer and hardware ownership. NVIDIA GeForce RTX 3090: 0.7171 h × 0.300 kW × €0.2896/kWh × 1.138 USD/EUR = ~$0.070898.

Top Models by Response Time (avg)

Instructions following Score vs Total Cost