modelbenchmark.io

Best AI for knowledge

Ranked by SimpleQA Verified, the one knowledge benchmark in the composite. Full leaderboard.

Job
View
RankModelLabCompositeEvidenceOut $/M
1Gemini 4 ArgonGoogle100th9 bench · 1 src$20.00
2GPT-6 AstraOpenAI99th15 bench · 2 src$50.00
3GPT-6.1 SolOpenAI98th13 bench · 2 src$10.00
4Gemini 3.1 Pro PreviewGoogle96th13 bench · 2 src$12.00
5Claude Opus 5.5Anthropic95th15 bench · 2 src$20.00
6Claude Fable 5.1Anthropic94th14 bench · 2 src$50.00
7Claude Fable 5Anthropic93rd15 bench · 2 src$50.00
8GPT-5.6 SolOpenAI92nd13 bench · 2 src$20.00
9Gemini 3.8 FlashGoogle90th9 bench · 2 src$7.50
10Gemini 3.7 FlashGoogle89th9 bench · 2 src$7.50
11Gemini 3 Flash PreviewGoogle88th10 bench · 2 src$3.00
12Gemini 3.6 FlashGoogle87th9 bench · 2 src$3.75
13Gemini 3.5 FlashGoogle85th14 bench · 3 src$9.00
14GPT-5.5OpenAI84th17 bench · 2 src$30.00
15GPT-6 SolOpenAI83rd13 bench · 2 src$10.00
16Muse Spark 1.2Meta82nd3 bench · 2 src$4.25
17Claude Opus 5Anthropic81st10 bench · 2 src$25.00
18Muse Spark 1.1Meta79th3 bench · 2 src$4.25
19Muse Spark OpenMeta78th5 bench · 1 src$4.25
20Grok 4.7xAI77th9 bench · 2 src$6.00
21Qwen3.7-MaxAlibaba76th10 bench · 2 src$7.50
22Claude Opus 4.8Anthropic74th14 bench · 2 src$25.00
23DeepSeek V4 Pro 0813DeepSeek73rd8 bench · 2 src$3.96
24Qwen3.6 Max PreviewAlibaba72nd6 bench · 1 srcfree
25Claude Opus 4.7Anthropic71st14 bench · 2 src$25.00
26Kimi K3Kimi70th9 bench · 2 src$1.00
27GPT 5OpenAI68th13 bench · 3 src$10.00
28o3OpenAI67th11 bench · 3 src$8.00
29Grok 4.6xAI66th10 bench · 2 src$6.00
30Qwen3 MaxAlibaba65th6 bench · 1 src$6.00
31Grok 4.5xAI63rd8 bench · 2 src$6.00
32GPT 5.1OpenAI62nd8 bench · 2 src$10.00
33Qwen3.8 Max (0902)Alibaba61st7 bench · 1 src—
34Claude Opus 4.6Anthropic60th12 bench · 2 src$25.00
35DeepSeek v4DeepSeek59th7 bench · 1 src—
36Claude Sonnet 5.5Anthropic57th13 bench · 2 src$10.00
37GPT-5.4 ProOpenAI56th6 bench · 1 src$180.00
38Qwen3.8 MaxAlibaba55th8 bench · 2 src$6.00
39Claude Opus 4.5Anthropic54th12 bench · 2 src$25.00
40GPT-5.4OpenAI52nd14 bench · 2 src$15.00
41Qwen3.6 PlusAlibaba51st9 bench · 2 src$3.00
42GPT-5.6 TerraOpenAI50th9 bench · 2 src$12.00
43DeepSeek V4.1 FlashDeepSeek49th9 bench · 2 src$1.20
44GPT-6 LunaOpenAI48th12 bench · 2 src$0.50
45o1-2024-12-17OpenAI46th9 bench · 3 src$54.00
46GPT-5.6 LunaOpenAI45th9 bench · 2 src$1.20
47GLM-5.3Zhipu44th8 bench · 2 src$4.40
48Qwen3-235B-A22BAlibaba43rd6 bench · 1 src—
49InklingThinking Machines42nd7 bench · 2 src$4.05
50Kimi K2.7 CodeKimi40th7 bench · 2 src$4.00
51Kimi K2.6Kimi39th13 bench · 2 src$4.00
52Kimi K2.5Kimi38th6 bench · 2 src$2.80
53GLM-5.2Zhipu37th10 bench · 2 src$4.40
54Claude Sonnet 4.6Anthropic35th8 bench · 2 src$15.94
55GLM-5.1Zhipu34th8 bench · 1 src$4.40
56DeepSeek V4 Flash 0731DeepSeek33rd8 bench · 2 src$0.28
57GPT 5.2OpenAI32nd13 bench · 3 src$14.00
58Claude Sonnet 5Anthropic31st8 bench · 2 src$10.00
59Grok 4.3 BetaxAI29th6 bench · 1 src—
60GLM-4.7Zhipu28th6 bench · 1 src$2.20
61GPT 4.1OpenAI27th10 bench · 3 src$8.00
62Claude Sonnet 4.5Anthropic26th10 bench · 1 src$15.00
63Grok 4.20xAI24th6 bench · 1 src$12.00
64GPT-5.4 miniOpenAI23rd10 bench · 2 src$4.50
65GPT 4oOpenAI22nd9 bench · 3 src$10.00
66Qwen3.5 PlusAlibaba21st7 bench · 1 src$2.21
67Claude Haiku 5.5Anthropic20th10 bench · 2 src$0.50
68Claude Sonnet 4.5Anthropic18th8 bench · 1 src—
69GPT 5 miniOpenAI17th11 bench · 2 src$2.00
70Qwen3.5 FlashAlibaba16th8 bench · 1 src$0.368
71Mistral Large 4Mistral15th9 bench · 2 src$4.4156
72o4-miniOpenAI13th12 bench · 3 src$4.00
73Inkling SmallThinking Machines12th6 bench · 1 src$1.20
74Qwen3.6 FlashAlibaba11th8 bench · 1 src$1.50
75O3 MiniOpenAI10th12 bench · 3 src$4.40
76Claude Haiku 4.5Anthropic9th7 bench · 1 src$5.00
77GPT 4.1 miniOpenAI7th10 bench · 3 src$1.60
78Claude 3 OpusAnthropic6th6 bench · 2 src$75.00
79GPT 5 nanoOpenAI5th11 bench · 2 src$0.40
80GPT-5.4 nanoOpenAI4th10 bench · 2 src$1.25
81Gemma 4 31B ITGoogle2nd3 bench · 1 srcfree
82GPT-4o miniOpenAI1st8 bench · 2 src$0.60
83GPT-4.1 nanoOpenAI0th7 bench · 2 src$0.40

Common questions

What is the best ai for knowledge right now?
Ranked by SimpleQA Verified, the one knowledge benchmark in the composite.
How is this ranking computed?
It is the same family composite as the leaderboard. Each benchmark is weighted by how much of its spread survives its own measurement error.