AI Model Leaderboard

Name: AllBenched AI model leaderboard
Creator: AllBenched

Reasoning ranking, Model families, sorted by Overall
Rank	Model	Provider	Weights	Overall	Coding	Agents	Reasoning	Math	Knowledge	Cost
30	DeepSeek V4.1 Flash 1017	DeepSeek	Open	60.9	62.9	61.6	63.4	60.7	55.4	$0.26
42	DeepSeek V4 Pro 0813	DeepSeek	Open	59.6	58.4	58.5	61.3	58.7	57.2	$0.99
44	DeepSeek V4.1 Flash	DeepSeek	Open	59.3	60.0	58.3	61.5	58.8	55.3	$0.33
62	DeepSeek V4 Flash 0731	DeepSeek	Open	57.8	55.2	59.1	60.1	57.5	53.1	$0.19
73	DeepSeek V4 Pro	DeepSeek	Open	55.9	54.5	56.9	58.0	54.0	54.4	$2.06
74	DeepSeek V4 Flash	DeepSeek	Open	55.9	50.7	–No results	56.0	55.7	–No results	$0.18
109	DeepSeek V3.2	DeepSeek	Open	50.9	52.5	50.4	50.5	51.5	50.8	$0.33
133	DeepSeek R1 0528	DeepSeek	Open	47.8	48.2	42.6	46.4	49.7	–No results	$1.01
146	DeepSeek-R1	DeepSeek	Open	46.6	45.3	43.3	45.9	46.6	48.0	$2.36
169	DeepSeek-R1-Distill-Llama-70B	DeepSeek	Open	42.9	–No results	–No results	41.0	45.2	–No results	–No price
176	DeepSeek V3 0324	DeepSeek	Open	41.7	44.7	41.0	41.7	41.2	42.6	$0.50
186	DeepSeek-R1-Distill-Qwen-14B	DeepSeek	Open	39.4	–No results	–No results	35.5	44.3	–No results	–No price
191	DeepSeek-V3	DeepSeek	Open	38.6	41.8	39.8	39.7	39.0	35.6	$0.51
238	DeepSeek LLM 67B	DeepSeek	Open	25.6	–No results	–No results	26.7	24.1	–No results	–No price