DeepSeekMath V2 Benchmark Scores & Performance

Benchmark analysis of DeepSeekMath V2 by DeepSeek across 14 tests.

Creator

DeepSeek

Source Type

Open Weight

Reasoning

Reasoning

Context Window

128K

Overall Score

64#36 of 88

Knowledge Benchmarks

MMLU
80
GPQA
79
SuperGPQA
77
OpenBookQA
75

Coding Benchmarks

HumanEval
72

Mathematics Benchmarks

AIME 2023
80
AIME 2024
82
AIME 2025
81
HMMT Feb 2023
76
HMMT Feb 2024
78
HMMT Feb 2025
77
BRUMO 2025
79

Reasoning Benchmarks

SimpleQA
77
MuSR
75

Frequently Asked Questions

How does DeepSeekMath V2 perform overall in AI benchmarks?

DeepSeekMath V2 ranks #36 out of 88 models with an overall score of 64. It is created by DeepSeek and features a 128K context window.

Is DeepSeekMath V2 good for knowledge and understanding?

DeepSeekMath V2 ranks #36 out of 88 models in knowledge and understanding benchmarks with an average score of 77.8. There are stronger options in this category.

Is DeepSeekMath V2 good for coding and programming?

DeepSeekMath V2 ranks #37 out of 88 models in coding and programming benchmarks with an average score of 72. There are stronger options in this category.

Is DeepSeekMath V2 good for mathematics?

DeepSeekMath V2 ranks #37 out of 88 models in mathematics benchmarks with an average score of 79. There are stronger options in this category.

Is DeepSeekMath V2 good for reasoning and logic?

DeepSeekMath V2 ranks #37 out of 88 models in reasoning and logic benchmarks with an average score of 76. There are stronger options in this category.

Is DeepSeekMath V2 open source?

Yes, DeepSeekMath V2 is an open weight model created by DeepSeek, meaning it can be downloaded and run locally or fine-tuned for specific use cases.

What is the context window size of DeepSeekMath V2?

DeepSeekMath V2 has a context window of 128K tokens, which determines how much text it can process in a single interaction.