Model Comparison

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

Gemma 3 1B

Google

Gemma 3 1B is a language model developed by Google. The model shows competitive results across 18 benchmarks. It excels particularly in IFEval (80.2%), GSM8k (62.8%), Natural2Code (56.0%). It's licensed for commercial use, making it suitable for enterprise applications. Released in 2025, it represents Google's latest advancement in AI technology.

Qwen2 72B Instruct

Alibaba

Qwen2 72B Instruct is a language model developed by Alibaba. It achieves strong performance with an average score of 73.6% across 17 benchmarks. It excels particularly in GSM8k (91.1%), CMMLU (90.1%), HellaSwag (87.6%). The model shows particular specialization in code tasks with an average performance of 82.3%. Released in 2024, it represents Alibaba's latest advancement in AI technology.

Qwen2 72B Instruct

Alibaba

2024-07-23

Gemma 3 1B

Google

2025-03-12

7 months newer

Average performance across 6 common benchmarks

Gemma 3 1B

Average Score:36.9%

Qwen2 72B Instruct

+33.7%

Average Score:70.6%

Performance comparison across key benchmark categories

Gemma 3 1B

code

43.0%

math

42.2%

general

17.8%

factuality

36.4%

Qwen2 72B Instruct

code

+39.3%

82.3%

math

+33.2%

75.4%

general

+50.1%

67.9%

factuality

+18.4%

54.8%

Benchmark Scores - Detailed View

Side-by-side comparison of all benchmark scores

Provider Availability & Performance

Available providers and their performance metrics

Gemma 3 1B

0 providers

Qwen2 72B Instruct

0 providers

Gemma 3 1B

Avg Score:36.9%

Providers:0

Qwen2 72B Instruct

+33.7%

Avg Score:70.6%

Providers:0