Puzzle Solving Model Ranking

AI BENCHY Category

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

Models Shown

Average Puzzle Solving Score

6.7

Best Model

Failure Reasons

With failure reason Wrong answer147 With failure reason Did not follow instructions78 With failure reason API error13 With failure reason Extra formatting5 With failure reason Timed out5 With failure reason No answer1

Rank	Model	Company	Puzzle Solving Score	Score	Tests Correct	Response Time (avg)
#123	MiMo-V2.5-Pro none	Xiaomi	6.7	5.5	1/3	1.30s
#117	Qwen3.5-35B-A3B none	Qwen	3.7	5.6	0/3	1.35s
#115	Qwen3.5-27B none	Qwen	6.7	5.7	1/3	1.38s
#149	Nemotron 3 Nano Omni 30b A3b Reasoning medium	NVIDIA	2.9	4.6	0/3	1.40s
#124	Kimi K2.6 none	Moonshot AI	3.1	5.5	0/3	1.40s
#61	Gemini 3.1 Flash Lite low	Google	10.0	7.2	3/3	1.40s
#125	GPT-5.4 none	OpenAI	5.6	5.5	1/3	1.44s
#32	Gemini 3.5 Flash minimal	Google	10.0	7.7	3/3	1.45s
#112	GLM 5.1 none	Z.ai	7.7	5.7	2/3	1.45s
#120	Mimo V2 PRO none	Xiaomi	6.0	5.6	1/3	1.61s
#50	Gemini 3.1 Flash Lite Preview low	Google	10.0	7.4	3/3	1.69s
#88	Qwen3.7 Plus none	Qwen	7.7	6.4	2/3	1.71s
#160	LFM2-24B-A2B none	Liquid	3.8	4.2	0/3	1.78s
#57	Step 3.7 Flash low	Stepfun	5.5	7.3	1/3	1.84s
#152	MiMo-V2-Flash none	Xiaomi	5.3	4.6	1/3	1.86s

Puzzle Solving Ranking