BENCHMARK EXPLORER · math

MATH Level 5

MATH Level 5 results published in Epoch AI’s benchmark data.

What it measures

Mean score in the named Epoch AI evaluation; reasoning configurations remain separate.

Interpretation caveats

Scores apply to the reported benchmark and exact model configuration. Other evaluators, dataset revisions and tool settings may differ.

Original benchmark

Family: math
Published versions: 1
Metric: Accuracy
Evidence: Source-reported

Published results

Version

Models ordered by MATH Level 5. Click a row to open its model profile.
#		Modalities
1	OpenAI	InputTextImageFileOutputText	400,000tokens
2	OpenAI	InputTextImageFileOutputText	400,000tokens
3	OpenAI	InputTextImageFileOutputText	400,000tokens
4	OpenAI	InputImageTextFileOutputText	200,000tokens
5	OpenAI	InputImageTextFileOutputText	200,000tokens
6	Anthropic	InputTextOutputText	—Not documented
7	Alibaba	InputTextOutputText	—Not documented
8	OpenAI	InputTextImageFileOutputText	400,000tokens
9	DeepSeek	InputTextOutputText	—Not documented
10	OpenAI	InputTextFileOutputText	200,000tokens
11	Anthropic	InputTextOutputText	—Not documented
12	Google DeepMind	InputTextOutputText	—Not documented
13	Google DeepMind	InputTextOutputText	—Not documented
14	OpenAI	InputTextImageFileOutputText	400,000tokens
15	OpenAI	InputTextFileOutputText	200,000tokens
16	OpenAI	InputTextImageFileOutputText	400,000tokens
17	OpenAI	InputTextImageFileOutputText	200,000tokens
18	OpenAI	InputTextImageFileOutputText	200,000tokens
19	DeepSeek	InputTextOutputText	—Not documented
20	Anthropic	InputTextOutputText	—Not documented
21	xAI	InputTextOutputText	—Not documented
22	Anthropic	InputTextOutputText	—Not documented
23	DeepSeek	InputTextOutputText	—Not documented
24	OpenAI	InputTextOutputText	—Not documented
25	xAI	InputTextOutputText	—Not documented
26	xAI	InputTextOutputText	—Not documented
27	OpenAI	InputImageTextFileOutputText	1,047,576tokens
28	DeepSeek	InputTextOutputText	—Not documented
29	Anthropic	InputTextOutputText	—Not documented
30	Anthropic	InputTextOutputText	—Not documented
31	Anthropic	InputTextOutputText	—Not documented
32	Anthropic	InputTextOutputText	—Not documented
33	OpenAI	InputTextOutputText	—Not documented
34	Google DeepMind	InputTextOutputText	—Not documented
35	OpenAI	InputImageTextFileOutputText	1,047,576tokens
36	Google DeepMind,Google	InputTextOutputText	—Not documented
37	OpenAI	InputTextOutputText	—Not documented
38	Mistral AI	InputTextOutputText	—Not documented
39	OpenAI	InputTextOutputText	—Not documented
40	DeepSeek	InputTextOutputText	—Not documented
41	Google DeepMind	InputTextOutputText	—Not documented
42	Meta AI	InputTextOutputText	—Not documented
43	Google DeepMind	InputTextOutputText	—Not documented
44	OpenAI	InputImageTextFileOutputText	1,047,576tokens
45	Alibaba	InputTextOutputText	—Not documented
46	Anthropic	InputTextOutputText	—Not documented
47	Alibaba	InputTextOutputText	—Not documented
48	Alibaba	InputTextOutputText	—Not documented
49	Microsoft Research	InputTextOutputText	16,384tokens
50	DeepSeek	InputTextOutputText	163,840tokens

Source-backed evidence. Different tasks. A clearer picture.

About Methodology Benchmark dashboard