Easy Benchmarks
Workspace
Overview
Benchmarks
Benchmarks list
Compare
Overall Index
Coding
Math
MMLU-Pro
Speed
Value
LLMs
Audio
Image
Video
Feedback
Log inSign up
Back

Artificial Analysis Math Index

Artificial Analysis aggregate math score.

The Math Index is the source dataset's aggregate math signal. It is useful for comparing models that need reliable numerical reasoning, but it should still be read next to direct math benchmarks such as MATH-500 and AIME.

Test type: Aggregate math evaluation that combines math-focused benchmark results exposed by Artificial Analysis.

Coverage

269 models have this metric.

99.0

Current leader: GPT-5.2 (xhigh)

Project links

Scores come from the Artificial Analysis LLM snapshot committed in this app.

Artificial Analysis methodology

Top Math Models

Top models ranked by Math.

Leaderboard

RankModelCreatorValueSpeedBlended Price
#1GPT-5.2 (xhigh)OpenAI99.071 tok/s$4.81/M
#2
GPT-5 Codex (high)
OpenAI
98.7
171.1 tok/s
$3.44/M
#3Gemini 3 Flash Preview (Reasoning)Google97.0172.8 tok/s$1.13/M
#4DeepSeek V3.2 SpecialeDeepSeek96.7n/a-
#5GPT-5.2 (medium)OpenAI96.7n/a$4.81/M
#6MiMo-V2-Flash (Reasoning)Xiaomi96.3129.5 tok/s$0.150/M
#7Gemini 3 Pro Preview (high)Google95.7n/a$4.50/M
#8GPT-5.1 Codex (high)OpenAI95.7182.1 tok/s$3.44/M
#9GLM-4.7 (Reasoning)Z AI95.079.2 tok/s$1.00/M
#10KAT-Coder-Pro V1KwaiKAT94.7114.7 tok/s$0.525/M
#11Kimi K2 ThinkingKimi94.7131.1 tok/s$1.08/M
#12GPT-5 (high)OpenAI94.3111.1 tok/s$3.44/M
#13Nova 2.0 Lite (high)Amazon94.3174.7 tok/s$0.850/M
#14GPT-5.1 (high)OpenAI94.0121.2 tok/s$3.44/M
#15gpt-oss-120b (high)OpenAI93.4348.5 tok/s$0.262/M
#16Grok 4xAI92.7n/a$11.00/M
#17DeepSeek V3.2 (Reasoning)DeepSeek92.0n/a$0.337/M
#18GPT-5 (medium)OpenAI91.785.6 tok/s$3.44/M
#19GPT-5.1 Codex mini (high)OpenAI91.7213.6 tok/s$0.688/M
#20Claude Opus 4.5 (Reasoning)Anthropic91.353.5 tok/s$10.94/M
#21NVIDIA Nemotron 3 Nano 30B A3B (Reasoning)NVIDIA91.0133.6 tok/s$0.096/M
#22Qwen3 235B A22B 2507 (Reasoning)Alibaba91.059.4 tok/s$0.838/M
#23GPT-5 mini (high)OpenAI90.787.4 tok/s$0.688/M
#24o4-mini (high)OpenAI90.7151 tok/s$1.93/M
#25K-EXAONE (Reasoning)LG AI Research90.3n/a-
#26DeepSeek V3.1 (Reasoning)DeepSeek89.7n/a$0.865/M
#27DeepSeek V3.1 Terminus (Reasoning)DeepSeek89.7n/a$1.91/M
#28Grok 4 Fast (Reasoning)xAI89.7n/a$0.275/M
#29Nova 2.0 Omni (medium)Amazon89.7n/a$0.850/M
#30gpt-oss-20B (high)OpenAI89.3240 tok/s$0.088/M
#31Grok 4.1 Fast (Reasoning)xAI89.3n/a-
#32Ring-1TInclusionAI89.3n/a-
#33Nova 2.0 Pro Preview (medium)Amazon89.0127.7 tok/s$3.44/M
#34Nova 2.0 Lite (medium)Amazon88.7190.6 tok/s$0.850/M
#35o3OpenAI88.3122.3 tok/s$3.50/M
#36Qwen3 VL 235B A22B (Reasoning)Alibaba88.332.5 tok/s$2.17/M
#37Apriel-v1.6-15B-ThinkerServiceNow88.0n/a-
#38Claude 4.5 Sonnet (Reasoning)Anthropic88.050.1 tok/s$6.56/M
#39INTELLECT-3Prime Intellect88.0n/a-
#40DeepSeek V3.2 Exp (Reasoning)DeepSeek87.7n/a$0.310/M
#41Gemini 2.5 ProGoogle87.7132 tok/s$3.44/M
#42Apriel-v1.5-15B-ThinkerServiceNow87.5n/a-
#43Gemini 3 Pro Preview (low)Google86.7n/a$4.50/M
#44GLM-4.6 (Reasoning)Z AI86.043.9 tok/s$0.963/M
#45GLM-4.6V (Reasoning)Z AI85.344.9 tok/s$0.450/M
#46ERNIE 5.0 Thinking PreviewBaidu85.0n/a-
#47GPT-5 mini (medium)OpenAI85.086.7 tok/s$0.688/M
#48Grok 3 mini Reasoning (high)xAI84.758.8 tok/s$0.350/M
#49Qwen3 VL 32B (Reasoning)Alibaba84.798.3 tok/s$2.63/M
#50Seed-OSS-36B-InstructByteDance Seed84.740.4 tok/s$0.300/M
#51Qwen3 Next 80B A3B (Reasoning)Alibaba84.3135.7 tok/s$1.88/M
#52Claude 4.5 Haiku (Reasoning)Anthropic83.7148.3 tok/s$2.00/M
#53GPT-5 nano (high)OpenAI83.7150.4 tok/s$0.138/M
#54Ring-flash-2.0InclusionAI83.7n/a$0.247/M
#55GPT-5 (low)OpenAI83.079.3 tok/s$3.44/M
#56MiniMax-M2.1MiniMax82.7184.6 tok/s$0.525/M
#57Qwen3 4B 2507 (Reasoning)Alibaba82.7n/a-
#58Qwen3 Max Thinking (Preview)Alibaba82.350.7 tok/s$2.40/M
#59Qwen3 VL 30B A3B (Reasoning)Alibaba82.3126.8 tok/s$0.338/M
#60Magistral Medium 1.2Mistral82.041.1 tok/s$2.75/M
#61Qwen3 235B A22B (Reasoning)Alibaba82.059 tok/s$2.63/M
#62GLM-4.5-AirZ AI80.774.5 tok/s$0.372/M
#63Qwen3 MaxAlibaba80.748.2 tok/s$3.05/M
#64Claude 4.1 Opus (Reasoning)Anthropic80.333.7 tok/s$32.81/M
#65Magistral Small 1.2Mistral80.3110.9 tok/s$0.750/M
#66Motif-2-12.7B-ReasoningMotif Technologies80.3n/a-
#67EXAONE 4.0 32B (Reasoning)LG AI Research80.0n/a-
#68Falcon-H1R-7BTII UAE80.0n/a-
#69Doubao Seed CodeByteDance Seed79.3n/a-
#70Mi:dm K 2.5 Pro PreviewKorea Telecom78.7n/a-
#71Gemini 2.5 Flash Preview (Sep '25) (Reasoning)Google78.3n/a-
#72GPT-5 nano (medium)OpenAI78.3167 tok/s$0.138/M
#73K2-V2 (high)MBZUAI Institute of Foundation Models78.3n/a-
#74MiniMax-M2MiniMax78.3102.9 tok/s$0.525/M
#75Olmo 3.1 32B ThinkAllen Institute for AI77.3n/a-
#76Llama Nemotron Super 49B v1.5 (Reasoning)NVIDIA76.744.2 tok/s$0.175/M
#77Mi:dm K 2.5 ProKorea Telecom76.7n/a-
#78DeepSeek R1 0528 (May '25)DeepSeek76.0n/a$2.06/M
#79NVIDIA Nemotron Nano 12B v2 VL (Reasoning)NVIDIA75.0n/a$0.300/M
#80Qwen3 Max (Preview)Alibaba75.047.1 tok/s$2.40/M