Model Comparison

Comprehensive side-by-side analysis of model capabilities and performance

DeepSeek R1 Zero

DeepSeek

DeepSeek R1 Zero is a language model developed by DeepSeek. It achieves strong performance with an average score of 76.5% across 4 benchmarks. It excels particularly in MATH-500 (95.9%), AIME 2024 (86.7%), GPQA (73.3%). It's licensed for commercial use, making it suitable for enterprise applications. Released in 2025, it represents DeepSeek's latest advancement in AI technology.

DeepSeek-V3.2-Exp

DeepSeek

DeepSeek-V3.2-Exp is a language model developed by DeepSeek. It achieves strong performance with an average score of 66.1% across 14 benchmarks. It excels particularly in SimpleQA (97.1%), AIME 2025 (89.3%), MMLU-Pro (85.0%). It supports a 229K token context window for handling large documents. The model is available through 2 API providers. It's licensed for commercial use, making it suitable for enterprise applications. Released in 2025, it represents DeepSeek's latest advancement in AI technology.

DeepSeek R1 Zero

DeepSeek

2025-01-20

DeepSeek-V3.2-Exp

DeepSeek

2025-09-29

8 months newer

Performance Metrics

Context window and performance specifications

DeepSeek R1 Zero

Max Context:-

Parameters:671.0B

DeepSeek-V3.2-Exp

Larger context

Max Context:229.4K

Parameters:685.0B

Average performance across 16 common benchmarks

DeepSeek R1 Zero

Average Score:19.1%

DeepSeek-V3.2-Exp

+38.7%

Average Score:57.8%

Performance comparison across key benchmark categories

DeepSeek R1 Zero

math

+9.4%

95.9%

general

+17.9%

80.0%

code

50.0%

DeepSeek-V3.2-Exp

math

86.5%

general

62.1%

code

+14.3%

64.3%

Provider Availability & Performance

Available providers and their performance metrics

DeepSeek R1 Zero

0 providers

DeepSeek-V3.2-Exp

2 providers

Novita

ZeroEval

Throughput: 100 tok/s

Latency: 0.7ms

DeepSeek R1 Zero

Avg Score:19.1%

Providers:0

DeepSeek-V3.2-Exp

+38.7%

Avg Score:57.8%

Providers:2