Skip to main content
benchlm.ai

Vals MATH 500 (MATH 500)

Academic math benchmark on probability, algebra, and trigonometry

Data verified

How BenchLM shows MATH 500

BenchLM mirrors the public Vals AI MATH 500 leaderboard captured from https://www.vals.ai/benchmarks/math500 and updated by Vals on January 9, 2026. The snapshot preserves overall scores, uncertainty, latency, cost-per-test metadata, and task-level scores where Vals publishes them.

MATH 500 is display only on BenchLM. Vals proprietary or Vals-hosted aggregate views are useful context, but BenchLM does not use them as weighted ranking inputs or as a replacement for benchmark-native source records.

60 Vals rows1 task viewspublic datasetTasks: OverallDisplay only
MATH 500 on Vals AIVals methodologyVals home

MATH 500 score on MATH 500 — January 9, 2026

BenchLM mirrors the published math 500 score view for MATH 500. Gemini 3 Pro Preview leads the public snapshot at 96.40% , followed by Grok 4 0709 (96.20%) and GPT-5 (96.00%). BenchLM does not use these results to rank models overall.

1

Gemini 3 Pro Preview

Google

google/gemini-3-pro-preview

96.40%
Overall —
2

Grok 4 0709

xAI

grok/grok-4-0709

96.20%
Overall —
3

GPT-5

OpenAI

openai/gpt-5-2025-08-07

96.00%
Overall —
60 modelsExternal benchmark mirrorsCurrentDisplay onlyUpdated January 9, 2026

The published MATH 500 snapshot is tightly clustered at the top: Gemini 3 Pro Preview sits at 96.40%, while the third row is only 0.40 points behind. The broader top-10 spread is 2.20 points, so many of the published scores sit in a relatively narrow band.

60 models have been evaluated on MATH 500. The benchmark falls in the External benchmark mirrors category. BenchLM tracks this category separately from its weighted global scoring system, so these results are best compared on the dedicated Korean benchmark views. MATH 500 is currently displayed for reference but excluded from the scoring formula, so it does not directly affect overall rankings.

About MATH 500

Year

2026

Tasks

MATH 500 academic math problems

Format

Accuracy score

Difficulty

Advanced academic math

BenchLM mirrors the public Vals AI MATH 500 leaderboard as display-only external evidence. The captured snapshot preserves overall scores, task-level scores where Vals publishes them, uncertainty, latency, and cost-per-test metadata. It is excluded from BenchLM weighted rankings.

MATH 500Public benchmark source

BenchLM freshness & provenance

Version

MATH 500 2026

Refresh cadence

Quarterly

Staleness state

Current

Question availability

Public benchmark set

CurrentDisplay only

BenchLM uses freshness metadata to decide whether a benchmark should still be treated as a strong differentiator, a benchmark to watch, or a display-only reference. For the full scoring policy, see the BenchLM methodology page.

MATH 500 score table (60 models)

1
Gemini 3 Pro Previewgoogle/gemini-3-pro-preview
96.40%
2
Grok 4 0709grok/grok-4-0709
96.20%
3
GPT-5openai/gpt-5-2025-08-07
96.00%
4
Claude Opus 4.1 20250805 Thinkinganthropic/claude-opus-4-1-20250805-thinking
95.40%
5
Gemini 2.5 Pro Exp 03 25google/gemini-2.5-pro-exp-03-25
95.20%
6
GPT-5 Miniopenai/gpt-5-mini-2025-08-07
94.80%
7
GPT Oss 120bfireworks/gpt-oss-120b
94.80%
8
O3openai/o3-2025-04-16
94.60%
9
Qwen3 235b A22bfireworks/qwen3-235b-a22b
94.60%
10
O4 Miniopenai/o4-mini-2025-04-16
94.20%
11
Grok 3 Mini Fast High Reasoninggrok/grok-3-mini-fast-high-reasoning
94.20%
12
GPT Oss 20bfireworks/gpt-oss-20b
94.20%
13
Moonshotai Kimi K2 Instructtogether/moonshotai/Kimi-K2-Instruct
94.20%
14
GLM 4.5zai/glm-4.5
94.00%
15
GPT-5 Nanoopenai/gpt-5-nano-2025-08-07
93.80%
16
Claude Sonnet 4 20250514 Thinkinganthropic/claude-sonnet-4-20250514-thinking
93.80%
17
Claude Opus 4.1anthropic/claude-opus-4-1-20250805
93.00%
18
DeepSeek R1fireworks/deepseek-r1
92.20%
19
O3 Miniopenai/o3-mini-2025-01-31
91.80%
20
Gemini 2.5 Flash Preview 04 17 Thinkinggoogle/gemini-2.5-flash-preview-04-17-thinking
91.80%
21
Gemini 2.5 Flash Preview 04 17google/gemini-2.5-flash-preview-04-17
91.60%
22
Claude 3.7 Sonnet 20250219 Thinkinganthropic/claude-3-7-sonnet-20250219-thinking
91.60%
23
Langston Nim Nvidia Llama 3.3 Nemotron Super 49b V1 42e84561 Thinkingtogether/langston/nim/nvidia/llama-3.3-nemotron-super-49b-v1-42e84561-thinking
91.40%
24
Claude Opus 4anthropic/claude-opus-4-20250514
90.40%
25
O1openai/o1-2024-12-17
90.40%
26
Claude Sonnet 4anthropic/claude-sonnet-4-20250514
90.32%
27
Grok 3grok/grok-3
89.80%
28
Gemini 2.0 Flash Expgoogle/gemini-2.0-flash-exp
89.00%
29
MiniMax M2.1minimax/MiniMax-M2.1
89.00%
30
DeepSeek V3 0324fireworks/deepseek-v3-0324
88.60%
31
Gemini 2.0 Flash 001google/gemini-2.0-flash-001
88.00%
32
GPT-4.1 Miniopenai/gpt-4.1-mini-2025-04-14
88.00%
33
GPT-4.1openai/gpt-4.1-2025-04-14
87.20%
34
Mistral Medium 2505mistralai/mistral-medium-2505
87.00%
35
Llama4 Maverick Instruct Basicfireworks/llama4-maverick-instruct-basic
85.20%
36
Gemini 2.0 Flash Thinking Exp 01 21google/gemini-2.0-flash-thinking-exp-01-21
84.60%
37
Gemini 1.5 Pro 002google/gemini-1.5-pro-002
82.80%
38
DeepSeek V3fireworks/deepseek-v3
80.40%
39
GPT-4.1 Nanoopenai/gpt-4.1-nano-2025-04-14
80.20%
40
Meta Llama Llama 4 Scout 17B 16E Instructtogether/meta-llama/Llama-4-Scout-17B-16E-Instruct
79.20%
41
Gemini 1.5 Flash 002google/gemini-1.5-flash-002
78.80%
42
Grok 2 1212grok/grok-2-1212
78.40%
43
Claude 3.7 Sonnetanthropic/claude-3-7-sonnet-20250219
76.80%
44
Command A 03 2025cohere/command-a-03-2025
76.20%
45
GPT-4oopenai/gpt-4o-2024-08-06
75.20%
46
Mistral Large 2411mistralai/mistral-large-2411
74.40%
47
GPT-4oopenai/gpt-4o-2024-11-20
74.00%
48
Meta Llama Llama 3.3 70B Instruct Turbotogether/meta-llama/Llama-3.3-70B-Instruct-Turbo
73.40%
49
GPT-4o Miniopenai/gpt-4o-mini-2024-07-18
72.60%
50
Claude 3.5 Sonnetanthropic/claude-3-5-sonnet-20241022
72.40%
51
Meta Llama Meta Llama 3.1 405B Instruct Turbotogether/meta-llama/Meta-Llama-3.1-405B-Instruct-Turbo
71.40%
52
Langston Nim Nvidia Llama 3.3 Nemotron Super 49b V1 42e84561together/langston/nim/nvidia/llama-3.3-nemotron-super-49b-v1-42e84561
71.20%
53
Mistral Small 2402mistralai/mistral-small-2402
70.60%
54
Grok 3 Mini Fast Low Reasoninggrok/grok-3-mini-fast-low-reasoning
70.20%
55
Mistral Small 2503mistralai/mistral-small-2503
68.40%
56
Meta Llama Meta Llama 3.1 70B Instruct Turbotogether/meta-llama/Meta-Llama-3.1-70B-Instruct-Turbo
65.00%
57
Claude 3.5 Haikuanthropic/claude-3-5-haiku-20241022
64.20%
58
Jamba Large 1.6ai21labs/jamba-large-1.6
54.80%
59
Meta Llama Meta Llama 3.1 8B Instruct Turbotogether/meta-llama/Meta-Llama-3.1-8B-Instruct-Turbo
44.40%
60
Jamba Mini 1.6ai21labs/jamba-mini-1.6
25.40%

FAQ

What does MATH 500 measure?

Academic math benchmark on probability, algebra, and trigonometry

Which model leads the published MATH 500 snapshot?

Gemini 3 Pro Preview currently leads the published MATH 500 snapshot with 96.40% math 500 score. BenchLM shows this benchmark for display only and does not use it in overall rankings.

How many models are evaluated on MATH 500?

60 AI models are included in BenchLM's mirrored MATH 500 snapshot, based on the public leaderboard captured on January 9, 2026.

Last updated: January 9, 2026 · mirrored from the public benchmark leaderboard

The AI models change fast. We track them for you.

For engineers, researchers, and the plain curious — a weekly brief on new models, ranking shifts, and pricing changes.

Free. No spam. Unsubscribe anytime.