GeekRadar / The comparison desk

COMPARE.

AI models

alibaba

Qwen2-72B-Instruct

Model evaluation

Capability & preference

Human preference · Arena

Overall text preferences from head-to-head votes. Not factual accuracy or a task-success rate.

Qwen2-72B-Instruct 1203.2 Qwen2-72B-Instruct · 95% interval 1198.3–1208.2 · 37,325 votes · 2026-10-28

Recorded results

Only the same benchmark, version, workload, scenario and unit share a row. Missing results are not zero. System results include the full submitted hardware and software configuration.

Arena text preference / chinesehuman preference · 2026-10-28 · rating

Qwen2-72B-Instruct

1,239.85 rating

Confidence interval: 1,228.28 – 1,251.42

4078 votes

Oct 28, 2026

Category: chinese
Evaluated configuration: Qwen2-72B-Instruct
Evaluation type: human preference

Arena text preference / codinghuman preference · 2026-10-28 · rating

Qwen2-72B-Instruct

1,195.54 rating

Confidence interval: 1,186.66 – 1,204.42

6249 votes

Oct 28, 2026

Category: coding
Evaluated configuration: Qwen2-72B-Instruct
Evaluation type: human preference

Arena text preference / creative_writinghuman preference · 2026-10-28 · rating

Qwen2-72B-Instruct

1,180.59 rating

Confidence interval: 1,171.12 – 1,190.07

5504 votes

Oct 28, 2026

Category: creative_writing
Evaluated configuration: Qwen2-72B-Instruct
Evaluation type: human preference

Arena text preference / englishhuman preference · 2026-10-28 · rating

Qwen2-72B-Instruct

1,222.07 rating

Confidence interval: 1,216.06 – 1,228.08

19928 votes

Oct 28, 2026

Category: english
Evaluated configuration: Qwen2-72B-Instruct
Evaluation type: human preference

Arena text preference / exclude_tieshuman preference · 2026-10-28 · rating

Qwen2-72B-Instruct

1,087.15 rating

Confidence interval: 1,079.83 – 1,094.48

23757 votes

Oct 28, 2026

Category: exclude_ties
Evaluated configuration: Qwen2-72B-Instruct
Evaluation type: human preference

Arena text preference / experthuman preference · 2026-10-28 · rating

Qwen2-72B-Instruct

1,170.97 rating

Confidence interval: 1,156.34 – 1,185.60

1763 votes

Oct 28, 2026

Category: expert
Evaluated configuration: Qwen2-72B-Instruct
Evaluation type: human preference

Arena text preference / frenchhuman preference · 2026-10-28 · rating

Qwen2-72B-Instruct

1,170.72 rating

Confidence interval: 1,144.11 – 1,197.33

596 votes

Oct 28, 2026

Category: french
Evaluated configuration: Qwen2-72B-Instruct
Evaluation type: human preference

Arena text preference / germanhuman preference · 2026-10-28 · rating

Qwen2-72B-Instruct

1,150.79 rating

Confidence interval: 1,132.77 – 1,168.80

1340 votes

Oct 28, 2026

Category: german
Evaluated configuration: Qwen2-72B-Instruct
Evaluation type: human preference

Arena text preference / hard_promptshuman preference · 2026-10-28 · rating

Qwen2-72B-Instruct

1,190.47 rating

Confidence interval: 1,183.20 – 1,197.74

10780 votes

Oct 28, 2026

Category: hard_prompts
Evaluated configuration: Qwen2-72B-Instruct
Evaluation type: human preference

Arena text preference / hard_prompts_englishhuman preference · 2026-10-28 · rating

Qwen2-72B-Instruct

1,205.75 rating

Confidence interval: 1,197.07 – 1,214.43

6483 votes

Oct 28, 2026

Category: hard_prompts_english
Evaluated configuration: Qwen2-72B-Instruct
Evaluation type: human preference

Arena text preference / industry_business_and_management_and_financial_operationshuman preference · 2026-10-28 · rating

Qwen2-72B-Instruct

1,146.31 rating

Confidence interval: 1,135.64 – 1,156.98

4345 votes

Oct 28, 2026

Category: industry_business_and_management_and_financial_operations
Evaluated configuration: Qwen2-72B-Instruct
Evaluation type: human preference

Arena text preference / industry_entertainment_and_sports_and_mediahuman preference · 2026-10-28 · rating

Qwen2-72B-Instruct

1,165.54 rating

Confidence interval: 1,156.41 – 1,174.67

5791 votes

Oct 28, 2026

Category: industry_entertainment_and_sports_and_media
Evaluated configuration: Qwen2-72B-Instruct
Evaluation type: human preference

Arena text preference / industry_legal_and_governmenthuman preference · 2026-10-28 · rating

Qwen2-72B-Instruct

1,193.17 rating

Confidence interval: 1,179.20 – 1,207.14

2262 votes

Oct 28, 2026

Category: industry_legal_and_government
Evaluated configuration: Qwen2-72B-Instruct
Evaluation type: human preference

Arena text preference / industry_life_and_physical_and_social_sciencehuman preference · 2026-10-28 · rating

Qwen2-72B-Instruct

1,195.19 rating

Confidence interval: 1,185.84 – 1,204.54

6456 votes

Oct 28, 2026

Category: industry_life_and_physical_and_social_science
Evaluated configuration: Qwen2-72B-Instruct
Evaluation type: human preference

Arena text preference / industry_mathematicalhuman preference · 2026-10-28 · rating

Qwen2-72B-Instruct

1,230.64 rating

Confidence interval: 1,220.67 – 1,240.61

4235 votes

Oct 28, 2026

Category: industry_mathematical
Evaluated configuration: Qwen2-72B-Instruct
Evaluation type: human preference

Arena text preference / industry_medicine_and_healthcarehuman preference · 2026-10-28 · rating

Qwen2-72B-Instruct

1,175.53 rating

Confidence interval: 1,160.96 – 1,190.09

2067 votes

Oct 28, 2026

Category: industry_medicine_and_healthcare
Evaluated configuration: Qwen2-72B-Instruct
Evaluation type: human preference

Arena text preference / industry_software_and_it_serviceshuman preference · 2026-10-28 · rating

Qwen2-72B-Instruct

1,192.85 rating

Confidence interval: 1,185.29 – 1,200.41

10308 votes

Oct 28, 2026

Category: industry_software_and_it_services
Evaluated configuration: Qwen2-72B-Instruct
Evaluation type: human preference

Arena text preference / industry_writing_and_literature_and_languagehuman preference · 2026-10-28 · rating

Qwen2-72B-Instruct

1,201.60 rating

Confidence interval: 1,193.80 – 1,209.41

9976 votes

Oct 28, 2026

Category: industry_writing_and_literature_and_language
Evaluated configuration: Qwen2-72B-Instruct
Evaluation type: human preference

Arena text preference / instruction_followinghuman preference · 2026-10-28 · rating

Qwen2-72B-Instruct

1,181.00 rating

Confidence interval: 1,174.24 – 1,187.76

14194 votes

Oct 28, 2026

Category: instruction_following
Evaluated configuration: Qwen2-72B-Instruct
Evaluation type: human preference

Arena text preference / japanesehuman preference · 2026-10-28 · rating

Qwen2-72B-Instruct

1,111.05 rating

Confidence interval: 1,089.73 – 1,132.38

901 votes

Oct 28, 2026

Category: japanese
Evaluated configuration: Qwen2-72B-Instruct
Evaluation type: human preference

Arena text preference / koreanhuman preference · 2026-10-28 · rating

Qwen2-72B-Instruct

1,082.65 rating

Confidence interval: 1,060.88 – 1,104.43

1167 votes

Oct 28, 2026

Category: korean
Evaluated configuration: Qwen2-72B-Instruct
Evaluation type: human preference

Arena text preference / longer_queryhuman preference · 2026-10-28 · rating

Qwen2-72B-Instruct

1,192.04 rating

Confidence interval: 1,181.91 – 1,202.18

4614 votes

Oct 28, 2026

Category: longer_query
Evaluated configuration: Qwen2-72B-Instruct
Evaluation type: human preference

Arena text preference / mathhuman preference · 2026-10-28 · rating

Qwen2-72B-Instruct

1,234.79 rating

Confidence interval: 1,225.38 – 1,244.19

4835 votes

Oct 28, 2026

Category: math
Evaluated configuration: Qwen2-72B-Instruct
Evaluation type: human preference

Arena text preference / multi_turnhuman preference · 2026-10-28 · rating

Qwen2-72B-Instruct

1,195.49 rating

Confidence interval: 1,186.30 – 1,204.67

6698 votes

Oct 28, 2026

Category: multi_turn
Evaluated configuration: Qwen2-72B-Instruct
Evaluation type: human preference

Arena text preference / non_englishhuman preference · 2026-10-28 · rating

Qwen2-72B-Instruct

1,175.77 rating

Confidence interval: 1,169.39 – 1,182.16

17397 votes

Oct 28, 2026

Category: non_english
Evaluated configuration: Qwen2-72B-Instruct
Evaluation type: human preference

Arena text preference / overallhuman preference · 2026-10-28 · rating

Qwen2-72B-Instruct

1,203.25 rating

Confidence interval: 1,198.35 – 1,208.15

37325 votes

Oct 28, 2026

Category: overall
Evaluated configuration: Qwen2-72B-Instruct
Evaluation type: human preference

Arena text preference / russianhuman preference · 2026-10-28 · rating

Qwen2-72B-Instruct

1,168.59 rating

Confidence interval: 1,158.19 – 1,178.98

4555 votes

Oct 28, 2026

Category: russian
Evaluated configuration: Qwen2-72B-Instruct
Evaluation type: human preference

Arena text preference / spanishhuman preference · 2026-10-28 · rating

Qwen2-72B-Instruct

1,168.86 rating

Confidence interval: 1,141.15 – 1,196.56

596 votes

Oct 28, 2026

Category: spanish
Evaluated configuration: Qwen2-72B-Instruct
Evaluation type: human preference

Run it locally

Inference · memory estimate

Qwen2-72B-Instruct

Not enough verified architecture information for a reliable VRAM estimate.

Capabilities

Model license: Qianwen LICENSE