GeekRadar / The comparison desk
COMPARE.
AI models
moonshot

Kimi K2 Preview (0905)

Model evaluation

Capability & preference

Human preference · Arena

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

Kimi K2 Preview (0905) 1379.9 Kimi K2 Preview (0905) · 95% interval 1373.4–1386.4 · 11,592 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

Kimi K2 Preview (0905)

1,415.25 rating

Confidence interval: 1,396.11 – 1,434.40

957 votes

Category
chinese
Evaluated configuration
Kimi K2 Preview (0905)
Evaluation type
human preference
Arena text preference / codinghuman preference · 2026-10-28 · rating

Kimi K2 Preview (0905)

1,399.09 rating

Confidence interval: 1,386.68 – 1,411.50

2232 votes

Category
coding
Evaluated configuration
Kimi K2 Preview (0905)
Evaluation type
human preference
Arena text preference / creative_writinghuman preference · 2026-10-28 · rating

Kimi K2 Preview (0905)

1,349.44 rating

Confidence interval: 1,334.22 – 1,364.67

1482 votes

Category
creative_writing
Evaluated configuration
Kimi K2 Preview (0905)
Evaluation type
human preference
Arena text preference / englishhuman preference · 2026-10-28 · rating

Kimi K2 Preview (0905)

1,378.70 rating

Confidence interval: 1,369.89 – 1,387.50

4781 votes

Category
english
Evaluated configuration
Kimi K2 Preview (0905)
Evaluation type
human preference
Arena text preference / exclude_tieshuman preference · 2026-10-28 · rating

Kimi K2 Preview (0905)

1,351.63 rating

Confidence interval: 1,343.14 – 1,360.12

8435 votes

Category
exclude_ties
Evaluated configuration
Kimi K2 Preview (0905)
Evaluation type
human preference
Arena text preference / experthuman preference · 2026-10-28 · rating

Kimi K2 Preview (0905)

1,365.25 rating

Confidence interval: 1,341.27 – 1,389.23

576 votes

Category
expert
Evaluated configuration
Kimi K2 Preview (0905)
Evaluation type
human preference
Arena text preference / frenchhuman preference · 2026-10-28 · rating

Kimi K2 Preview (0905)

1,380.93 rating

Confidence interval: 1,342.30 – 1,419.56

216 votes

Category
french
Evaluated configuration
Kimi K2 Preview (0905)
Evaluation type
human preference
Arena text preference / germanhuman preference · 2026-10-28 · rating

Kimi K2 Preview (0905)

1,386.66 rating

Confidence interval: 1,351.60 – 1,421.71

277 votes

Category
german
Evaluated configuration
Kimi K2 Preview (0905)
Evaluation type
human preference
Arena text preference / hard_promptshuman preference · 2026-10-28 · rating

Kimi K2 Preview (0905)

1,383.60 rating

Confidence interval: 1,375.26 – 1,391.94

5556 votes

Category
hard_prompts
Evaluated configuration
Kimi K2 Preview (0905)
Evaluation type
human preference
Arena text preference / hard_prompts_englishhuman preference · 2026-10-28 · rating

Kimi K2 Preview (0905)

1,378.38 rating

Confidence interval: 1,366.31 – 1,390.44

2392 votes

Category
hard_prompts_english
Evaluated configuration
Kimi K2 Preview (0905)
Evaluation type
human preference
Arena text preference / industry_business_and_management_and_financial_operationshuman preference · 2026-10-28 · rating

Kimi K2 Preview (0905)

1,353.26 rating

Confidence interval: 1,340.13 – 1,366.40

2048 votes

Category
industry_business_and_management_and_financial_operations
Evaluated configuration
Kimi K2 Preview (0905)
Evaluation type
human preference
Arena text preference / industry_entertainment_and_sports_and_mediahuman preference · 2026-10-28 · rating

Kimi K2 Preview (0905)

1,341.66 rating

Confidence interval: 1,328.85 – 1,354.46

2079 votes

Category
industry_entertainment_and_sports_and_media
Evaluated configuration
Kimi K2 Preview (0905)
Evaluation type
human preference
Arena text preference / industry_legal_and_governmenthuman preference · 2026-10-28 · rating

Kimi K2 Preview (0905)

1,402.64 rating

Confidence interval: 1,381.62 – 1,423.65

768 votes

Category
industry_legal_and_government
Evaluated configuration
Kimi K2 Preview (0905)
Evaluation type
human preference
Arena text preference / industry_life_and_physical_and_social_sciencehuman preference · 2026-10-28 · rating

Kimi K2 Preview (0905)

1,386.78 rating

Confidence interval: 1,372.53 – 1,401.03

1740 votes

Category
industry_life_and_physical_and_social_science
Evaluated configuration
Kimi K2 Preview (0905)
Evaluation type
human preference
Arena text preference / industry_mathematicalhuman preference · 2026-10-28 · rating

Kimi K2 Preview (0905)

1,399.43 rating

Confidence interval: 1,376.84 – 1,422.03

641 votes

Category
industry_mathematical
Evaluated configuration
Kimi K2 Preview (0905)
Evaluation type
human preference
Arena text preference / industry_medicine_and_healthcarehuman preference · 2026-10-28 · rating

Kimi K2 Preview (0905)

1,382.65 rating

Confidence interval: 1,360.12 – 1,405.19

719 votes

Category
industry_medicine_and_healthcare
Evaluated configuration
Kimi K2 Preview (0905)
Evaluation type
human preference
Arena text preference / industry_software_and_it_serviceshuman preference · 2026-10-28 · rating

Kimi K2 Preview (0905)

1,393.83 rating

Confidence interval: 1,384.12 – 1,403.53

3860 votes

Category
industry_software_and_it_services
Evaluated configuration
Kimi K2 Preview (0905)
Evaluation type
human preference
Arena text preference / industry_writing_and_literature_and_languagehuman preference · 2026-10-28 · rating

Kimi K2 Preview (0905)

1,345.49 rating

Confidence interval: 1,333.84 – 1,357.14

2579 votes

Category
industry_writing_and_literature_and_language
Evaluated configuration
Kimi K2 Preview (0905)
Evaluation type
human preference
Arena text preference / instruction_followinghuman preference · 2026-10-28 · rating

Kimi K2 Preview (0905)

1,347.58 rating

Confidence interval: 1,336.60 – 1,358.55

2889 votes

Category
instruction_following
Evaluated configuration
Kimi K2 Preview (0905)
Evaluation type
human preference
Arena text preference / japanesehuman preference · 2026-10-28 · rating

Kimi K2 Preview (0905)

1,348.69 rating

Confidence interval: 1,308.69 – 1,388.68

209 votes

Category
japanese
Evaluated configuration
Kimi K2 Preview (0905)
Evaluation type
human preference
Arena text preference / koreanhuman preference · 2026-10-28 · rating

Kimi K2 Preview (0905)

1,322.70 rating

Confidence interval: 1,288.03 – 1,357.37

313 votes

Category
korean
Evaluated configuration
Kimi K2 Preview (0905)
Evaluation type
human preference
Arena text preference / longer_queryhuman preference · 2026-10-28 · rating

Kimi K2 Preview (0905)

1,352.84 rating

Confidence interval: 1,341.37 – 1,364.30

2578 votes

Category
longer_query
Evaluated configuration
Kimi K2 Preview (0905)
Evaluation type
human preference
Arena text preference / mathhuman preference · 2026-10-28 · rating

Kimi K2 Preview (0905)

1,397.30 rating

Confidence interval: 1,376.66 – 1,417.94

755 votes

Category
math
Evaluated configuration
Kimi K2 Preview (0905)
Evaluation type
human preference
Arena text preference / multi_turnhuman preference · 2026-10-28 · rating

Kimi K2 Preview (0905)

1,365.14 rating

Confidence interval: 1,352.04 – 1,378.23

2030 votes

Category
multi_turn
Evaluated configuration
Kimi K2 Preview (0905)
Evaluation type
human preference
Arena text preference / non_englishhuman preference · 2026-10-28 · rating

Kimi K2 Preview (0905)

1,372.43 rating

Confidence interval: 1,364.67 – 1,380.19

6811 votes

Category
non_english
Evaluated configuration
Kimi K2 Preview (0905)
Evaluation type
human preference
Arena text preference / overallhuman preference · 2026-10-28 · rating

Kimi K2 Preview (0905)

1,379.88 rating

Confidence interval: 1,373.39 – 1,386.37

11592 votes

Category
overall
Evaluated configuration
Kimi K2 Preview (0905)
Evaluation type
human preference
Arena text preference / polishhuman preference · 2026-10-28 · rating

Kimi K2 Preview (0905)

1,377.67 rating

Confidence interval: 1,355.03 – 1,400.31

680 votes

Category
polish
Evaluated configuration
Kimi K2 Preview (0905)
Evaluation type
human preference
Arena text preference / russianhuman preference · 2026-10-28 · rating

Kimi K2 Preview (0905)

1,384.27 rating

Confidence interval: 1,364.31 – 1,404.22

839 votes

Category
russian
Evaluated configuration
Kimi K2 Preview (0905)
Evaluation type
human preference
Arena text preference / spanishhuman preference · 2026-10-28 · rating

Kimi K2 Preview (0905)

1,387.45 rating

Confidence interval: 1,356.18 – 1,418.72

390 votes

Category
spanish
Evaluated configuration
Kimi K2 Preview (0905)
Evaluation type
human preference

Run it locally

Inference · memory estimate

Kimi K2 Preview (0905)

Not enough verified architecture information for a reliable VRAM estimate.

Capabilities

Model license
Modified MIT