Puzzle Solving Model Ranking

See which AI models perform best on Puzzle Solving, which ones stay reliable, and where the biggest gaps appear. Sort by: Tests Correct ↓.

Models Shown

Average Puzzle Solving Score

0.7

Best Model

Gemini 3.6 Flash 1.0

Failure Reasons

With failure reason Wrong answer344 With failure reason Did not follow instructions160 With failure reason API error35 With failure reason Extra formatting16 With failure reason No answer14 With failure reason Timed out11

442/442

Rank	Model	Company	Puzzle Solving Score	Score	Total Cost	Tests Correct	Response Time (avg)
#304	Qwen3.5-35B-A3B none	Qwen	0.4	0.6	$0.120	0/3	1.35s
Total Tests 3 Wrong Tests 3 Total Cost $0.120 Response Time (avg) 1.35s
#305	Nemotron 3.5 Super VL none	NVIDIA	0.5	0.6	$0.040	0/3	754ms
Total Tests 3 Wrong Tests 3 Total Cost $0.040 Response Time (avg) 754ms
#306	Nemotron 3 Super medium	NVIDIA	0.3	0.6	$0.081	0/3	3.15s
Total Tests 3 Wrong Tests 3 Total Cost $0.081 Response Time (avg) 3.15s
#307	Kimi K2.6 none	Moonshot AI	0.3	0.5	$0.486	0/3	1.40s
Total Tests 3 Wrong Tests 3 Total Cost $0.486 Response Time (avg) 1.40s
#316	North Mini Code medium	Cohere	0.3	0.5	$0.000	0/3	19.7s
Total Tests 3 Wrong Tests 3 Total Cost $0.000 Response Time (avg) 19.7s
#317	Hy4 preview low	Tencent	0.5	0.5	$1.844	0/3	238.0s
Total Tests 3 Wrong Tests 3 Total Cost $1.844 Response Time (avg) 238.0s
#318	MiMo-V2.6-Flash none	Xiaomi	0.4	0.5	$0.060	0/3	739ms
Total Tests 3 Wrong Tests 3 Total Cost $0.060 Response Time (avg) 739ms
#323	Granite 4.2 8B low	IBM Granite	0.3	0.5	$0.036	0/3	90.8s
Total Tests 3 Wrong Tests 3 Total Cost $0.036 Response Time (avg) 90.8s
#324	Kimi K2.5 none	Moonshot AI	0.3	0.5	$0.262	0/3	4.04s
Total Tests 3 Wrong Tests 3 Total Cost $0.262 Response Time (avg) 4.04s
#326	Laguna XS 2.1 none	Poolside	0.3	0.5	$0.020	0/3	1.01s
Total Tests 3 Wrong Tests 3 Total Cost $0.020 Response Time (avg) 1.01s
#327	Qwen3.6 35B A3B none	Qwen	0.3	0.5	$0.105	0/3	1.07s
Total Tests 3 Wrong Tests 3 Total Cost $0.105 Response Time (avg) 1.07s
#330	Nemotron 3.5 Lightning high	NVIDIA	0.4	0.5	$0.168	0/3	6.09s
Total Tests 3 Wrong Tests 3 Total Cost $0.168 Response Time (avg) 6.09s
#331	KAT Coder AIR V2.5 high	Kwaipilot	0.4	0.5	$0.077	0/3	2.47s
Total Tests 3 Wrong Tests 3 Total Cost $0.077 Response Time (avg) 2.47s
#334	Mistral Small 4 medium	Mistral	0.3	0.5	$0.126	0/3	2.17s
Total Tests 3 Wrong Tests 3 Total Cost $0.126 Response Time (avg) 2.17s
#335	Mistral Small 4 none	Mistral	0.3	0.5	$0.047	0/3	399ms
Total Tests 3 Wrong Tests 3 Total Cost $0.047 Response Time (avg) 399ms

Puzzle Solving Ranking

Filter models

Top Models by Puzzle Solving Score

Puzzle Solving Score vs Total Cost

Top Models by Response Time (avg)