AgMoDB
ModelsAgentsOrchestratorsEvalsCompositesVisualizeIndustry
AgMoDB by @mistakeknot
All Benchmarks

Math Index

Artificial Analysis's composite score for mathematical ability, combining MATH-500, AIME, and other math benchmarks.

mathSource

Strengths

Aggregates several math benchmarks to reduce variance from any one test set; helps rank models when individual math leaderboards disagree.

Caveats

Weighting is set externally, so shifts in component emphasis can move rankings without true capability change; composite obscures which specific math domains drive wins or regressions.

How to interpret scores

Contamination risk: moderateFreshness: continuous

Higher is better. Composite math score — use for relative ranking between models.

Relevant use cases

Math assistanceQuantitative financeScientific computing

Caveat to keep in mind

Weighting is set externally, so shifts in component emphasis can move rankings without true capability change.

Leaderboard

Last synced: Nov 22, 2026, 12:18 PM
#ModelProviderScorePrice/M
1
Z.ai: GLM 4.5
source artificial_analysis
Evidence observed Nov 22, 2026
Z AI73.7$0.65
2
Qwen: Qwen3 30B A3B Instruct 2507
source artificial_analysis
Evidence observed Nov 22, 2026
Alibaba66.3$0.14
3
Google: Gemini 3 Flash Preview
source artificial_analysis
Evidence observed Nov 22, 2026
Google55.7$1.13
4
Anthropic: Claude Haiku 4.5
source artificial_analysis
Evidence observed Nov 22, 2026
Anthropic39.0$2.00
5
Anthropic: Claude Sonnet 4
source artificial_analysis
Evidence observed Oct 15, 2026
Anthropic38.0$6.00
6
Anthropic: Claude Sonnet 4.5
source artificial_analysis
Evidence observed Nov 22, 2026
Anthropic37.0$6.00
7
Anthropic: Claude Opus 4
source artificial_analysis
Evidence observed Sep 21, 2026
Anthropic36.3$30.00
8
DeepSeek: DeepSeek V3
source artificial_analysis
Evidence observed Nov 22, 2026
DeepSeek26.0$0.52
9
NVIDIA: Llama 3.1 Nemotron 70B Instruct
source artificial_analysis
Evidence observed May 7, 2026
NVIDIA11.0$1.20
10
Meta: Llama 3.3 70B Instruct
source artificial_analysis
Evidence observed Nov 22, 2026
Meta7.7$0.15
11
Meta: Llama 3.3 70B Instruct (free)
source artificial_analysis
Evidence observed Jul 10, 2026
Meta7.7$0.00
12
Meta: Llama 3.1 8B Instruct
source artificial_analysis
Evidence observed Nov 22, 2026
Meta4.3$0.03
13
Meta: Llama 3.1 70B Instruct
source artificial_analysis
Evidence observed Nov 22, 2026
Meta4.0$0.40
14
Meta: Llama 3.2 3B Instruct
source artificial_analysis
Evidence observed Nov 22, 2026
Meta3.3$0.02
15
Meta: Llama 3.2 3B Instruct (free)
source artificial_analysis
Evidence observed Jul 10, 2026
Meta3.3$0.00
16
AI21: Jamba Large 1.7
source artificial_analysis
Evidence observed Aug 19, 2026
AI21 Labs2.3$3.50
17
Meta: Llama 3.2 11B Vision Instruct
source artificial_analysis
Evidence observed Jul 10, 2026
Meta1.7$0.05
18
Meta: Llama 3.2 1B Instruct
source artificial_analysis
Evidence observed Nov 20, 2026
Meta0.0$0.07