MATH-500 Leaderboard

Competition mathematics problems requiring multi-step reasoning, covering algebra, geometry, number theory, and calculus.

Data from Artificial Analysis

As of November 21, 2026, the top-scoring model on MATH-500 is GPT-5 at 99.4%, followed by o3 at 99.2% and Grok 3 Mini at 99.2%. 112 models have been evaluated on this benchmark.

Last updated: November 21, 2026

8 Ways to Use Fewer Tokens

Models

112

Best Score

99.4

Average

82.2

Std Dev

16.9

Categories
Mathematical Problem Solving
Sponsored The Grid | Spot Priced LLM API
The Grid | Spot Priced LLM API — Why pay list price when providers will bid for your API usage? Up to 80% off.
View spot prices
Provider
Model
Input $/M
Output $/M
MATH-500
Actions
$1.250
$10.000
99.4
$2.000
$8.000
99.2
$0.250
$0.500
99.2
$3.000
$15.000
99.1
$3.000
$15.000
99.0
$1.100
$4.400
98.9
$0.625
$5.000
98.6
$1.100
$4.400
98.5
$0.450
$0.900
98.4
$0.000
$0.000
98.3
$0.500
$2.150
98.3
$15.000
$75.000
98.2
$0.300
$2.500
98.1
$1.250
$10.000
98.0
$0.000
$0.000
98.0
$0.000
$0.000
98.0
$0.600
$2.200
97.9
$0.200
$0.900
97.6
$0.090
$0.300
97.5
$1.100
$4.400
97.3
$0.550
$2.200
97.1
$15.000
$60.000
97.0
$0.100
$0.400
96.9
$1.000
$10.000
96.7
$0.550
$2.190
96.6
$0.130
$0.850
96.5
$0.080
$0.200
96.1
$0.064
$0.224
96.1
$0.120
$0.500
95.9
$0.900
$0.900
95.7
$2.000
$8.000
95.7
$3.000
$15.000
94.7
$0.250
$1.000
94.2
$0.300
$0.900
94.2
$0.300
$0.300
94.1
$15.000
$75.000
94.1
$0.750
$0.800
93.5
$3.000
$15.000
93.4
$0.200
$0.200
93.3
$0.300
$2.500
93.2
$0.900
$0.900
93.1
$0.150
$0.600
93.0
$0.900
$0.900
93.0
$0.100
$0.400
92.6
$0.400
$1.600
92.5
$2.000
$8.000
91.3
$0.440
$2.200
90.7
$0.200
$0.200
90.4
$0.900
$0.900
90.2
$0.070
$0.280
89.3
$0.200
$0.600
88.9
$0.080
$0.160
88.3
$0.075
$0.200
88.3
$0.075
$0.300
87.3
$0.080
$0.200
87.1
$3.000
$15.000
87.0
$0.064
$0.224
86.9
$0.120
$0.500
86.3
$0.288
$0.320
85.8
$0.050
$0.150
85.3
$3.000
$15.000
85.0
$0.100
$0.400
84.8
$0.100
$0.300
84.4
$0.200
$0.200
84.3
$2.500
$12.500
83.9
$1.040
$4.160
83.5
$0.200
$0.200
82.8
$2.500
$10.000
81.9
$1.000
$1.000
81.7
$0.070
$0.140
81.0
$0.033
$0.130
80.5
$0.900
$0.900
80.5
$2.500
$10.000
79.5
$0.150
$0.600
78.9
$0.800
$3.200
78.6
$0.000
$0.000
77.5
$0.100
$0.320
77.3
$3.000
$15.000
77.1
$0.000
$0.000
77.1
$0.660
$0.800
76.7
$0.040
$0.050
76.6
$0.060
$0.240
76.5
$3.000
$15.000
74.5
$5.000
$15.000
73.7
$2.000
$6.000
73.6
$0.880
$0.880
73.3
$0.800
$4.000
72.1
$0.040
$0.064
71.5
$2.000
$6.000
71.4
$2.000
$6.000
71.4
$0.400
$2.000
70.7
$0.100
$0.300
70.7
$1.500
$1.500
70.3
$0.035
$0.140
70.3
$0.200
$0.600
67.7
$0.200
$0.200
66.0
$0.340
$0.390
64.9
$0.070
$0.280
63.5
$2.000
$8.000
60.0
$30.000
$60.000
56.8
$1.200
$1.200
54.5
$0.100
$0.280
53.8
$0.500
$1.500
52.7
$0.020
$0.040
51.9
$0.060
$0.060
51.6
$0.040
$0.040
49.9
$0.030
$0.050
48.9
$0.510
$0.740
48.3
$0.500
$1.000
44.1
$0.250
$1.250
39.4
$0.140
$0.420
29.9
$0.020
$0.020
14.0

Pricing from OpenRouter. Benchmarks from Artificial Analysis.

About MATH-500

Competition mathematics problems requiring multi-step reasoning, covering algebra, geometry, number theory, and calculus.

This leaderboard shows all models with MATH-500 benchmark scores, ranked from highest to lowest. Pricing data is included to help you compare performance against cost.

Frequently Asked Questions

Competition mathematics problems requiring multi-step reasoning, covering algebra, geometry, number theory, and calculus.
As of November 21, 2026, GPT-5 leads the MATH-500 leaderboard with a score of 99.4. Rankings change as new models are released and evaluated.
Currently 112 models have been evaluated on MATH-500, with an average score of 82.2 and standard deviation of 16.9.
Benchmark scores are updated when new evaluations are published by our data sources (Artificial Analysis and LayerLens). Pricing data is refreshed daily from OpenRouter.