GeekRadar / The comparison desk
COMPARE.
AI models
moonshot

Kimi K2.5 (Instant)

Model evaluation

Capability & preference

Human preference · Arena

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

Kimi K2.5 (Instant) 1420.1 Kimi K2.5 (Instant) · 95% interval 1413.7–1426.6 · 8,371 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 K2.5 (Instant)

1,450.63 rating

Confidence interval: 1,426.04 – 1,475.22

543 votes

Category
chinese
Evaluated configuration
Kimi K2.5 (Instant)
Evaluation type
human preference
Arena text preference / codinghuman preference · 2026-10-28 · rating

Kimi K2.5 (Instant)

1,483.47 rating

Confidence interval: 1,470.09 – 1,496.84

1868 votes

Category
coding
Evaluated configuration
Kimi K2.5 (Instant)
Evaluation type
human preference
Arena text preference / creative_writinghuman preference · 2026-10-28 · rating

Kimi K2.5 (Instant)

1,382.03 rating

Confidence interval: 1,366.07 – 1,397.99

1324 votes

Category
creative_writing
Evaluated configuration
Kimi K2.5 (Instant)
Evaluation type
human preference
Arena text preference / englishhuman preference · 2026-10-28 · rating

Kimi K2.5 (Instant)

1,432.64 rating

Confidence interval: 1,422.53 – 1,442.75

3245 votes

Category
english
Evaluated configuration
Kimi K2.5 (Instant)
Evaluation type
human preference
Arena text preference / exclude_tieshuman preference · 2026-10-28 · rating

Kimi K2.5 (Instant)

1,409.04 rating

Confidence interval: 1,399.79 – 1,418.29

5775 votes

Category
exclude_ties
Evaluated configuration
Kimi K2.5 (Instant)
Evaluation type
human preference
Arena text preference / experthuman preference · 2026-10-28 · rating

Kimi K2.5 (Instant)

1,439.79 rating

Confidence interval: 1,417.20 – 1,462.38

632 votes

Category
expert
Evaluated configuration
Kimi K2.5 (Instant)
Evaluation type
human preference
Arena text preference / frenchhuman preference · 2026-10-28 · rating

Kimi K2.5 (Instant)

1,403.75 rating

Confidence interval: 1,357.97 – 1,449.54

158 votes

Category
french
Evaluated configuration
Kimi K2.5 (Instant)
Evaluation type
human preference
Arena text preference / germanhuman preference · 2026-10-28 · rating

Kimi K2.5 (Instant)

1,412.88 rating

Confidence interval: 1,366.79 – 1,458.97

144 votes

Category
german
Evaluated configuration
Kimi K2.5 (Instant)
Evaluation type
human preference
Arena text preference / hard_promptshuman preference · 2026-10-28 · rating

Kimi K2.5 (Instant)

1,443.11 rating

Confidence interval: 1,434.73 – 1,451.49

4726 votes

Category
hard_prompts
Evaluated configuration
Kimi K2.5 (Instant)
Evaluation type
human preference
Arena text preference / hard_prompts_englishhuman preference · 2026-10-28 · rating

Kimi K2.5 (Instant)

1,449.33 rating

Confidence interval: 1,435.73 – 1,462.92

1798 votes

Category
hard_prompts_english
Evaluated configuration
Kimi K2.5 (Instant)
Evaluation type
human preference
Arena text preference / industry_business_and_management_and_financial_operationshuman preference · 2026-10-28 · rating

Kimi K2.5 (Instant)

1,404.68 rating

Confidence interval: 1,390.09 – 1,419.28

1544 votes

Category
industry_business_and_management_and_financial_operations
Evaluated configuration
Kimi K2.5 (Instant)
Evaluation type
human preference
Arena text preference / industry_entertainment_and_sports_and_mediahuman preference · 2026-10-28 · rating

Kimi K2.5 (Instant)

1,396.24 rating

Confidence interval: 1,381.48 – 1,411.01

1494 votes

Category
industry_entertainment_and_sports_and_media
Evaluated configuration
Kimi K2.5 (Instant)
Evaluation type
human preference
Arena text preference / industry_legal_and_governmenthuman preference · 2026-10-28 · rating

Kimi K2.5 (Instant)

1,425.25 rating

Confidence interval: 1,400.47 – 1,450.03

532 votes

Category
industry_legal_and_government
Evaluated configuration
Kimi K2.5 (Instant)
Evaluation type
human preference
Arena text preference / industry_life_and_physical_and_social_sciencehuman preference · 2026-10-28 · rating

Kimi K2.5 (Instant)

1,433.65 rating

Confidence interval: 1,417.37 – 1,449.93

1263 votes

Category
industry_life_and_physical_and_social_science
Evaluated configuration
Kimi K2.5 (Instant)
Evaluation type
human preference
Arena text preference / industry_mathematicalhuman preference · 2026-10-28 · rating

Kimi K2.5 (Instant)

1,437.72 rating

Confidence interval: 1,408.16 – 1,467.27

361 votes

Category
industry_mathematical
Evaluated configuration
Kimi K2.5 (Instant)
Evaluation type
human preference
Arena text preference / industry_medicine_and_healthcarehuman preference · 2026-10-28 · rating

Kimi K2.5 (Instant)

1,416.12 rating

Confidence interval: 1,391.88 – 1,440.36

571 votes

Category
industry_medicine_and_healthcare
Evaluated configuration
Kimi K2.5 (Instant)
Evaluation type
human preference
Arena text preference / industry_software_and_it_serviceshuman preference · 2026-10-28 · rating

Kimi K2.5 (Instant)

1,470.50 rating

Confidence interval: 1,459.77 – 1,481.23

2882 votes

Category
industry_software_and_it_services
Evaluated configuration
Kimi K2.5 (Instant)
Evaluation type
human preference
Arena text preference / industry_writing_and_literature_and_languagehuman preference · 2026-10-28 · rating

Kimi K2.5 (Instant)

1,397.22 rating

Confidence interval: 1,384.16 – 1,410.28

1947 votes

Category
industry_writing_and_literature_and_language
Evaluated configuration
Kimi K2.5 (Instant)
Evaluation type
human preference
Arena text preference / instruction_followinghuman preference · 2026-10-28 · rating

Kimi K2.5 (Instant)

1,429.81 rating

Confidence interval: 1,418.03 – 1,441.59

2343 votes

Category
instruction_following
Evaluated configuration
Kimi K2.5 (Instant)
Evaluation type
human preference
Arena text preference / koreanhuman preference · 2026-10-28 · rating

Kimi K2.5 (Instant)

1,377.97 rating

Confidence interval: 1,330.77 – 1,425.17

138 votes

Category
korean
Evaluated configuration
Kimi K2.5 (Instant)
Evaluation type
human preference
Arena text preference / longer_queryhuman preference · 2026-10-28 · rating

Kimi K2.5 (Instant)

1,435.48 rating

Confidence interval: 1,423.75 – 1,447.20

2386 votes

Category
longer_query
Evaluated configuration
Kimi K2.5 (Instant)
Evaluation type
human preference
Arena text preference / mathhuman preference · 2026-10-28 · rating

Kimi K2.5 (Instant)

1,441.76 rating

Confidence interval: 1,417.91 – 1,465.62

556 votes

Category
math
Evaluated configuration
Kimi K2.5 (Instant)
Evaluation type
human preference
Arena text preference / multi_turnhuman preference · 2026-10-28 · rating

Kimi K2.5 (Instant)

1,427.43 rating

Confidence interval: 1,412.47 – 1,442.39

1496 votes

Category
multi_turn
Evaluated configuration
Kimi K2.5 (Instant)
Evaluation type
human preference
Arena text preference / non_englishhuman preference · 2026-10-28 · rating

Kimi K2.5 (Instant)

1,406.35 rating

Confidence interval: 1,398.31 – 1,414.38

5126 votes

Category
non_english
Evaluated configuration
Kimi K2.5 (Instant)
Evaluation type
human preference
Arena text preference / overallhuman preference · 2026-10-28 · rating

Kimi K2.5 (Instant)

1,420.13 rating

Confidence interval: 1,413.67 – 1,426.59

8371 votes

Category
overall
Evaluated configuration
Kimi K2.5 (Instant)
Evaluation type
human preference
Arena text preference / polishhuman preference · 2026-10-28 · rating

Kimi K2.5 (Instant)

1,379.81 rating

Confidence interval: 1,337.48 – 1,422.14

176 votes

Category
polish
Evaluated configuration
Kimi K2.5 (Instant)
Evaluation type
human preference
Arena text preference / russianhuman preference · 2026-10-28 · rating

Kimi K2.5 (Instant)

1,403.69 rating

Confidence interval: 1,387.02 – 1,420.36

1131 votes

Category
russian
Evaluated configuration
Kimi K2.5 (Instant)
Evaluation type
human preference
Arena text preference / spanishhuman preference · 2026-10-28 · rating

Kimi K2.5 (Instant)

1,446.57 rating

Confidence interval: 1,415.08 – 1,478.05

347 votes

Category
spanish
Evaluated configuration
Kimi K2.5 (Instant)
Evaluation type
human preference

Run it locally

Inference · memory estimate

Kimi K2.5 (Instant)

Not enough verified architecture information for a reliable VRAM estimate.

Capabilities

Model license
Modified MIT