Qwen2.5 72B Instruct

Qwen

Qwen2.5-72B-Instruct is a 72 billion parameter language model from the Qwen2.5 series, fine-tuned for instruction following. This model excels at producing extended text outputs exceeding 8,000 tokens, demonstrates strong comprehension of structured information like tables, and is adept at generating structured outputs, with a particular focus on JSON format. Furthermore, it boasts multilingual support for more than 29 languages.

Model Specifications

Technical details and capabilities of Qwen2.5 72B Instruct

Core Specifications

72.7B Parameters

Model size and complexity

18000.0B Training Tokens

Amount of data used in training

131.1K / 8.2K

Input / Output tokens

September 18, 2024

Release date

Capabilities & License

Multimodal Support

Not Supported

Web Hydrated

License

qwen

Resources

API Reference

https://www.alibabacloud.com/help/en/model-studio/developer-reference/use-qwen-by-calling-api

Code Repository

https://github.com/QwenLM/Qwen2.5

Performance Insights

Check out how Qwen2.5 72B Instruct handles various AI tasks through comprehensive benchmark results.

100

95.8

GSM8K

95.8

(96%)

93.5

MT-bench

93.5

(94%)

88.2

MBPP

88.2

(88%)

86.8

MMLU-Redux

86.8

(87%)

86.6

HumanEval

86.6

(87%)

84.1

IFEval

84.1

(84%)

83.1

MATH

83.1

(83%)

81.6

AlignBench

81.6

(82%)

81.2

Arena-Hard

81.2

(81%)

75.1

MultiPL-E

75.1

(75%)

71.1

MMLU-Pro

71.1

(71%)

55.5

LiveCodeBench

55.5

(56%)

52.3

LiveBench

52.3

(52%)

GPQA

(49%)

GSM8K

MT-bench

MBPP

MMLU-Redux

HumanEval

IFEval

MATH

AlignBench

Arena-Hard

MultiPL-E

MMLU-Pro

LiveCodeBench

LiveBench

GPQA

Model Comparison

See how Qwen2.5 72B Instruct stacks up against other leading models across key performance metrics.

100

71.1

MMLU-Pro - Qwen2.5 72B Instruct

71.1

(71%)

MMLU-Pro - Qwen2.5 32B Instruct

(69%)

63.7

MMLU-Pro - Qwen2.5 14B Instruct

63.7

(64%)

73.3

MMLU-Pro - Llama 3.1 405B Instruct

73.3

(73%)

75.8

MMLU-Pro - Gemini 1.5 Pro

75.8

(76%)

67.3

MMLU-Pro - Gemini 1.5 Flash

67.3

(67%)

GPQA - Qwen2.5 72B Instruct

(49%)

49.5

GPQA - Qwen2.5 32B Instruct

49.5

(50%)

45.5

GPQA - Qwen2.5 14B Instruct

45.5

(46%)

50.7

GPQA - Llama 3.1 405B Instruct

50.7

(51%)

59.1

GPQA - Gemini 1.5 Pro

59.1

(59%)

GPQA - Gemini 1.5 Flash

(51%)

83.1

MATH - Qwen2.5 72B Instruct

83.1

(83%)

83.1

MATH - Qwen2.5 32B Instruct

83.1

(83%)

MATH - Qwen2.5 14B Instruct

(80%)

73.8

MATH - Llama 3.1 405B Instruct

73.8

(74%)

86.5

MATH - Gemini 1.5 Pro

86.5

(87%)

77.9

MATH - Gemini 1.5 Flash

77.9

(78%)

95.8

GSM8K - Qwen2.5 72B Instruct

95.8

(96%)

95.9

GSM8K - Qwen2.5 32B Instruct

95.9

(96%)

94.8

GSM8K - Qwen2.5 14B Instruct

94.8

(95%)

96.8

GSM8K - Llama 3.1 405B Instruct

96.8

(97%)

90.8

GSM8K - Gemini 1.5 Pro

90.8

(91%)

86.2

GSM8K - Gemini 1.5 Flash

86.2

(86%)

86.6

HumanEval - Qwen2.5 72B Instruct

86.6

(87%)

88.4

HumanEval - Qwen2.5 32B Instruct

88.4

(88%)

83.5

HumanEval - Qwen2.5 14B Instruct

83.5

(84%)

HumanEval - Llama 3.1 405B Instruct

(89%)

84.1

HumanEval - Gemini 1.5 Pro

84.1

(84%)

74.3

HumanEval - Gemini 1.5 Flash

74.3

(74%)

MMLU-Pro

GPQA

MATH

GSM8K

HumanEval

Qwen2.5 72B Instruct

Qwen2.5 32B Instruct

Qwen2.5 14B Instruct

Llama 3.1 405B Instruct

Gemini 1.5 Pro

Gemini 1.5 Flash

Detailed Benchmarks

Dive deeper into Qwen2.5 72B Instruct's performance across specific task categories. Expand each section to see detailed metrics and comparisons.

Math

GSM8K

97.1%

96.4%

96.4%

95.9%

95.9%

95.8%

95.1%

95.0%

94.8%

68.6%

Current model

Other models

Avg (93.1%)

Coding

HumanEval

Claude 3.5 Sonnet

93.7%

88.1%

87.8%

87.2%

87.1%

86.6%

86.0%

85.7%

85.4%

Ministral 8B Instruct

34.8%

Current model

Other models

Avg (82.2%)

MBPP

Qwen2.5-Coder 32B Instruct

90.2%

Qwen2.5 72B Instruct

88.2%

Qwen2.5 32B Instruct

84.0%

Qwen2.5-Coder 7B Instruct

83.5%

82.0%

80.8%

80.2%

79.2%

Current model

Other models

Avg (83.5%)

LiveCodeBench

Grok-3 Mini

80.0%

Gemini Pro 2.5 Experimental

70.4%

65.9%

63.4%

62.5%

55.5%

50.0%

41.8%

40.5%

Qwen2.5-Coder 7B Instruct

18.2%

Current model

Other models

Avg (54.8%)

Knowledge

GPQA

87.7%

Llama 3.1 405B Instruct

50.7%

Llama 3.3 70B Instruct

50.5%

50.4%

49.5%

49.0%

48.0%

46.9%

Llama 3.2 90B Instruct

46.7%

Qwen2 7B Instruct

25.3%

Current model

Other models

Avg (50.5%)

MATH

97.9%

89.0%

86.5%

85.5%

83.1%

83.1%

80.4%

80.0%

78.3%

32.6%

Current model

Other models

Avg (79.6%)

Non categorized

MMLU-Pro

DeepSeek-R1

84.0%

GPT-4o

74.7%

Llama 3.1 405B Instruct

73.3%

72.6%

72.0%

71.1%

70.4%

69.0%

Llama 3.3 70B Instruct

68.9%

Qwen2.5-Coder 7B Instruct

40.1%

Current model

Other models

Avg (69.6%)

MMLU-Redux

92.9%

89.1%

86.8%

83.9%

80.0%

Qwen2.5-Coder 32B Instruct

77.5%

Qwen2.5 7B Instruct

75.4%

Qwen2.5-Coder 7B Instruct

66.6%

Current model

Other models

Avg (81.5%)

MultiPL-E

75.4%

75.1%

72.8%

70.4%

69.2%

59.1%

Current model

Other models

Avg (70.3%)

LiveBench

QwQ 32B

73.1%

67.0%

52.3%

52.3%

47.6%

35.9%

Current model

Other models

Avg (54.7%)

IFEval

Claude 3.7 Sonnet

93.2%

Llama 3.1 70B Instruct

87.5%

87.2%

86.1%

85.6%

84.1%

84.0%

83.9%

83.3%

61.3%

Current model

Other models

Avg (83.6%)

Arena-Hard

Qwen2.5 72B Instruct

81.2%

Qwen2.5 7B Instruct

52.0%

Current model

Other models

Avg (66.6%)

AlignBench

81.6%

80.4%

73.3%

72.1%

Current model

Other models

Avg (76.9%)

MT-bench

Qwen2.5 72B Instruct

93.5%

Qwen2.5 7B Instruct

87.5%

Current model

Other models

Avg (90.5%)

Providers Pricing Coming Soon

We're working on gathering comprehensive pricing data from all major providers for Qwen2.5 72B Instruct. Compare costs across platforms to find the best pricing for your use case.

OpenAI

Anthropic

Google

Mistral AI

Cohere

Share your feedback

Hi, I'm Charlie Palars, the founder of Deepranking.ai. I'm always looking for ways to improve the site and make it more useful for you. You can write me through this form or directly through X at @palarsio.

Your feedback helps us improve our service

Qwen2.5 72B Instruct

Model Specifications

Core Specifications

Capabilities & License

Resources

Performance Insights

Model Comparison

Detailed Benchmarks

Math

GSM8K

Coding

HumanEval

MBPP

LiveCodeBench

Knowledge

GPQA

MATH

Non categorized

MMLU-Pro

MMLU-Redux

MultiPL-E

LiveBench

IFEval

Arena-Hard

AlignBench

MT-bench

Providers Pricing Coming Soon

Share your feedback

Stay Ahead with AI Updates