Skip to content
VECTOR WIREAI INTELLIGENCE
UTC

MATH leaderboard — the family board.

7 TRACKED VARIANTS · 3 FIRST-CLASS BOARDS

One family, 7 tracked prints: MATH Lvl 5 (HF Open LLM) and MATH Maj@1 are cited standalone and keep live boards of their own. Scoring configs and splits of the base board render as sections of this page, each under its own honest label.

● LIVE DATA97 MODELS ON THIS BOARDLEADER DEEPSEEK-R1 AT 97.3

First-class boards in this familyCITED STANDALONE · EACH KEEPS ITS OWN LIVE BOARD

#BoardModels scored
01MATHTHIS PAGE97

The board — MATHTOP 97 OF 97 MEASURED · BEST SCORE PER MODEL, DEFAULT VARIANT

#ModelVendorScoreVLast seen
01DeepSeek-R1DeepSeek97.32026-05-01
02o1OpenAI96.42026-05-01
03Gemini 1.5 ProGoogle92.02026-09-07
04Kimi VL A3B ThinkingMoonshot91.82026-06-05
05DeepSeek-V3DeepSeek91.22026-06-06
06Gemini 1.5 Flash (001)Google90.82026-05-10
07Gemini 2.0 Flash ExpGoogle90.12026-06-06
08o1-miniOpenAI902026-05-01
09MAI-Thinking-1Microsoft89.52026-08-14
10Gemma 3 27BGoogle892026-08-25
11Qwen2.5 Instruct 72BAlibaba88.42026-06-06
12Claude 3.5 HaikuAnthropic87.22026-08-25
13MiMo V2.5 Pro BaseXiaomi86.22026-06-05
14GPT-4 TurboOpenAI85.72026-06-18
15GPT-4oOpenAI85.32026-09-07
16Qwen2.5 Instruct Turbo (7B)Alibaba83.52026-05-10
17Llama 3.1 Instruct 405BMeta82.72026-08-29
18Qwen2.5 VL 32B InstructAlibaba82.22026-08-25
19Nova ProAmazon82.12026-05-10
21Claude 3.5 SonnetAnthropic81.32026-09-07
22Llama 3.3 70B InstructMeta80.82026-05-30
23Phi 4Microsoft80.42026-05-30
24GPT-4OpenAI80.22026-08-12
25GPT-4o miniOpenAI80.22026-08-25
27Qwen2 Instruct (72B)Alibaba79.02026-05-10
28Llama 3.1 70B InstructMeta78.32026-05-31
29Nova LiteAmazon77.92026-06-23
30Hunyuan Large InstructTencent77.42026-05-31
31MiniMax Text 01MiniMax77.42026-06-06
32Palmyra-X-004Writer76.72026-05-10
33Hy3-preview-BaseTencent76.32026-06-05
34Claude 3 OpusAnthropic76.02026-09-07
35Nova MicroAmazon762026-05-10
36Qwen 2.5 14BAlibaba75.62026-05-30
37Mistral LargeMistral75.02026-05-10
38DeepSeek-V2.5DeepSeek74.72026-05-31
39Gemma 2 27B ITGoogle74.62026-05-10
41Gemma 2 Instruct (9B)Google72.42026-05-10
42Palmyra X V3 (72B)Writer72.32026-05-10
43Kimi K2 BaseMoonshot71.22026-06-13
44Yi Large (Preview)01.AI71.22026-05-10
45MiMo V2 Flash BaseXiaomi712026-06-13
46Phi 3 Small 8K InstructMicrosoft70.32026-05-10
47Llama 3.1 8B InstructMeta70.32026-05-10
48Hunyuan LargeTencent69.82026-05-31
49Jamba 1.5 LargeAI2169.22026-05-10
50MiMo V2.5 BaseXiaomi67.72026-06-05
51Mistral Large 2Mistral67.72026-05-10
52Mistral NemoMistral66.82026-05-10
53Step3.5 Flash BaseStepFun66.82026-06-05
54Llama 3 (70B)Meta66.32026-09-10
55DeepSeek-V4-Pro-BaseDeepSeek64.52026-06-05
56Mistral Large 3 675B Base 2512Mistral62.92026-06-05
57GLM-4.5-BaseZ.ai62.62026-06-05
58DeepSeek-V3.1-BaseDeepSeek62.62026-06-13
59DeepSeek-V3.2-Exp-BaseDeepSeek62.52026-06-13
60Mistral Small (2402)Mistral62.02026-05-10
61DeepSeek-LLM-67B-Chat (V1)DeepSeek61.52026-05-10
62Phi 3 Medium 4K InstructMicrosoft61.12026-09-21
63DeepSeek-V3-BaseDeepSeek59.42026-06-05
64Palmyra X V2 (33B)Writer58.02026-05-10
65DeepSeek-V4-Flash-BaseDeepSeek57.42026-06-05
66Solar ProUpstage56.72026-05-10
67Mistral Medium (Non-Reasoning)Mistral56.52026-09-21
68Arctic InstructSnowflake51.92026-05-10
69Qwen2 72BAlibaba51.12026-09-10
70Phi 3 Mini 4K InstructMicrosoft50.62026-08-29
71Gemma (7B)Google50.02026-08-29
72MoonlightMoonshot45.32026-06-12
73DeepSeek-V2DeepSeek43.62026-09-10
74Claude 3 SonnetAnthropic43.12026-09-07
75Qwen2.5 3BAlibaba42.62026-06-05
76Llama 3 Instruct 8BMeta41.12026-08-29
77Command R+ (Apr '24)Cohere40.32026-05-10
78Jamba InstructAI2138.02026-05-10
79DBRX InstructDatabricks35.82026-05-10
80Jamba 1.5 MiniAI2131.82026-05-10
81Mistral V0.1 (7B)Mistral29.72026-05-10
82Mistral Instruct V0.3 (7B)Mistral28.92026-05-10
83Command RCohere26.62026-05-10
84Llama (65B)Meta25.72026-05-10
85Phi 2Microsoft25.52026-05-10
86Mistral 7BMistral25.42026-08-29
87CommandCohere23.62026-05-10
88Inflection-1Inflection AI162026-08-12
89Claude 3 HaikuAnthropic13.12026-05-10
90Jurassic 2 Jumbo (178B)AI2110.32026-05-10
91Command LightCohere9.82026-05-10
92Llama 3.2 3B InstructMeta8.52026-06-05
93Luminous Supreme (70B)Aleph Alpha7.82026-05-10
94Jurassic 2 Grande (17B)AI216.42026-05-10
95Luminous Extended (30B)Aleph Alpha4.02026-05-10
96OLMo (7B)AllenAI2.92026-05-10
97Luminous Base (13B)Aleph Alpha2.62026-05-10
◆ IV — INDEPENDENTLY VERIFIED · ◆ AA — AGGREGATOR · ◆ VA — VENDOR · ◇ SA — SOURCE-ATTRIBUTED

Variants of this boardSCORING CONFIGS & SPLITS — NOT CITED STANDALONE, SO THEY LIVE HERE

MATH-500 (EM)

VARIANT64 models · best score per model
#ModelVendorScoreVLast seen
01Gemini 2.5 ProGoogle98.82026-06-05
02Sarvam 105BSarvam AI98.62026-08-23
03GLM-4.5Z.ai98.22026-08-23
04Claude Opus 4Anthropic98.22026-06-15
05o3OpenAI98.12026-06-05
06GLM-4.5-AirZ.ai98.12026-08-23
07NVIDIA Nemotron Nano 9B V2NVIDIA97.82026-08-23
08Kimi K2 InstructMoonshot97.42026-08-23
09DeepSeek-R1DeepSeek97.32026-09-11
10Sarvam 30BSarvam AI972026-08-23
12LongCat Flash LiteMeituan96.82026-08-23
13MiniMax M1 80KMiniMax96.82026-08-23
14Qwen3 14BAlibaba96.82026-09-09
15o1OpenAI96.42026-09-11

MATH (EM)

VARIANT11 models · best score per model
#ModelVendorScoreVLast seen
01Qwen2.5 7B BaseAlibaba85.42026-05-03
02Qwen3 4B Base DenseAlibaba75.42026-05-03
03DeepSeek-V4-Pro-BaseDeepSeek64.52026-09-10
04DeepSeek-V3DeepSeek61.62026-07-06
05DeepSeek-V3.2-BaseDeepSeek60.52026-06-27
06DeepSeek-V4-Flash-BaseDeepSeek57.42026-09-10
07Qwen2.5 Instruct 72BAlibaba54.42026-07-06
08Llama 3.1 Instruct 405BMeta492026-07-06
09DeepSeek-V2DeepSeek43.42026-07-06
10Qwen2.5 3B BaseAlibaba17.72026-05-03
11Gemma 3 1B Base DenseGoogle7.22026-05-03

MATH CoT 2-Shot

VARIANT6 models · best score per model
#ModelVendorScoreVLast seen
01Ministral 3 14BMistral67.62026-06-12
02Ministral 3 8BMistral62.62026-06-05
03Qwen3 14B BaseAlibaba622026-06-05
04Ministral 3 3BMistral60.12026-06-05
05Qwen3 8B BaseAlibaba57.62026-06-05
06Qwen3 4B BaseAlibaba40.52026-06-05

Other tracked prints

THIN BOARDSunder 5 published models each — listed, never hidden
PrintLeaderScoreModels
THE FAMILY BOARD · EVERY PRINT LABELED · EVERY SCORE SOURCE-ATTRIBUTED ON ITS OWN LIVE BOARDAll leaderboardsThe frontier boardModels by vendor