GeekRadar / The comparison desk
COMPARE.
AI models
xai

grok-4.20-beta1

Model evaluation

Capability & preference

Human preference · Arena

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

grok-4.20-beta1 1444.6 grok-4.20-beta1 · 95% interval 1440.1–1449.1 · 27,908 votes · 2026-10-15

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 · 2026-10-15 · rating

grok-4.20-beta1

1,476.38 rating

Confidence interval: 1,461.84 – 1,490.92

1738 votes

Category
chinese
Evaluated configuration
grok-4.20-beta1
Evaluation type
human preference
Arena text preference / codinghuman preference · 2026-10-15 · rating

grok-4.20-beta1

1,446.64 rating

Confidence interval: 1,439.15 – 1,454.13

7499 votes

Category
coding
Evaluated configuration
grok-4.20-beta1
Evaluation type
human preference
Arena text preference / creative_writinghuman preference · 2026-10-15 · rating

grok-4.20-beta1

1,437.87 rating

Confidence interval: 1,428.26 – 1,447.48

4173 votes

Category
creative_writing
Evaluated configuration
grok-4.20-beta1
Evaluation type
human preference
Arena text preference / englishhuman preference · 2026-10-15 · rating

grok-4.20-beta1

1,450.99 rating

Confidence interval: 1,444.95 – 1,457.03

12796 votes

Category
english
Evaluated configuration
grok-4.20-beta1
Evaluation type
human preference
Arena text preference / exclude_tieshuman preference · 2026-10-15 · rating

grok-4.20-beta1

1,440.68 rating

Confidence interval: 1,434.83 – 1,446.54

21178 votes

Category
exclude_ties
Evaluated configuration
grok-4.20-beta1
Evaluation type
human preference
Arena text preference / experthuman preference · 2026-10-15 · rating

grok-4.20-beta1

1,427.57 rating

Confidence interval: 1,415.29 – 1,439.84

2473 votes

Category
expert
Evaluated configuration
grok-4.20-beta1
Evaluation type
human preference
Arena text preference / frenchhuman preference · 2026-10-15 · rating

grok-4.20-beta1

1,455.23 rating

Confidence interval: 1,433.81 – 1,476.65

861 votes

Category
french
Evaluated configuration
grok-4.20-beta1
Evaluation type
human preference
Arena text preference / germanhuman preference · 2026-10-15 · rating

grok-4.20-beta1

1,464.10 rating

Confidence interval: 1,436.82 – 1,491.38

445 votes

Category
german
Evaluated configuration
grok-4.20-beta1
Evaluation type
human preference
Arena text preference / hard_promptshuman preference · 2026-10-15 · rating

grok-4.20-beta1

1,441.01 rating

Confidence interval: 1,435.58 – 1,446.45

17459 votes

Category
hard_prompts
Evaluated configuration
grok-4.20-beta1
Evaluation type
human preference
Arena text preference / hard_prompts_englishhuman preference · 2026-10-15 · rating

grok-4.20-beta1

1,443.64 rating

Confidence interval: 1,436.48 – 1,450.79

8403 votes

Category
hard_prompts_english
Evaluated configuration
grok-4.20-beta1
Evaluation type
human preference
Arena text preference / industry_business_and_management_and_financial_operationshuman preference · 2026-10-15 · rating

grok-4.20-beta1

1,425.00 rating

Confidence interval: 1,416.51 – 1,433.49

5360 votes

Category
industry_business_and_management_and_financial_operations
Evaluated configuration
grok-4.20-beta1
Evaluation type
human preference
Arena text preference / industry_entertainment_and_sports_and_mediahuman preference · 2026-10-15 · rating

grok-4.20-beta1

1,417.67 rating

Confidence interval: 1,409.12 – 1,426.23

5387 votes

Category
industry_entertainment_and_sports_and_media
Evaluated configuration
grok-4.20-beta1
Evaluation type
human preference
Arena text preference / industry_legal_and_governmenthuman preference · 2026-10-15 · rating

grok-4.20-beta1

1,433.26 rating

Confidence interval: 1,419.91 – 1,446.61

2107 votes

Category
industry_legal_and_government
Evaluated configuration
grok-4.20-beta1
Evaluation type
human preference
Arena text preference / industry_life_and_physical_and_social_sciencehuman preference · 2026-10-15 · rating

grok-4.20-beta1

1,440.82 rating

Confidence interval: 1,431.76 – 1,449.88

4607 votes

Category
industry_life_and_physical_and_social_science
Evaluated configuration
grok-4.20-beta1
Evaluation type
human preference
Arena text preference / industry_mathematicalhuman preference · 2026-10-15 · rating

grok-4.20-beta1

1,433.75 rating

Confidence interval: 1,418.36 – 1,449.14

1521 votes

Category
industry_mathematical
Evaluated configuration
grok-4.20-beta1
Evaluation type
human preference
Arena text preference / industry_medicine_and_healthcarehuman preference · 2026-10-15 · rating

grok-4.20-beta1

1,450.91 rating

Confidence interval: 1,437.25 – 1,464.57

2018 votes

Category
industry_medicine_and_healthcare
Evaluated configuration
grok-4.20-beta1
Evaluation type
human preference
Arena text preference / industry_software_and_it_serviceshuman preference · 2026-10-15 · rating

grok-4.20-beta1

1,452.45 rating

Confidence interval: 1,446.05 – 1,458.85

10842 votes

Category
industry_software_and_it_services
Evaluated configuration
grok-4.20-beta1
Evaluation type
human preference
Arena text preference / industry_writing_and_literature_and_languagehuman preference · 2026-10-15 · rating

grok-4.20-beta1

1,427.98 rating

Confidence interval: 1,420.12 – 1,435.85

6385 votes

Category
industry_writing_and_literature_and_language
Evaluated configuration
grok-4.20-beta1
Evaluation type
human preference
Arena text preference / instruction_followinghuman preference · 2026-10-15 · rating

grok-4.20-beta1

1,415.13 rating

Confidence interval: 1,408.19 – 1,422.08

8847 votes

Category
instruction_following
Evaluated configuration
grok-4.20-beta1
Evaluation type
human preference
Arena text preference / japanesehuman preference · 2026-10-15 · rating

grok-4.20-beta1

1,448.91 rating

Confidence interval: 1,408.94 – 1,488.88

226 votes

Category
japanese
Evaluated configuration
grok-4.20-beta1
Evaluation type
human preference
Arena text preference / koreanhuman preference · 2026-10-15 · rating

grok-4.20-beta1

1,416.72 rating

Confidence interval: 1,388.76 – 1,444.68

469 votes

Category
korean
Evaluated configuration
grok-4.20-beta1
Evaluation type
human preference
Arena text preference / longer_queryhuman preference · 2026-10-15 · rating

grok-4.20-beta1

1,426.88 rating

Confidence interval: 1,420.13 – 1,433.63

10449 votes

Category
longer_query
Evaluated configuration
grok-4.20-beta1
Evaluation type
human preference
Arena text preference / mathhuman preference · 2026-10-15 · rating

grok-4.20-beta1

1,432.54 rating

Confidence interval: 1,418.20 – 1,446.88

1684 votes

Category
math
Evaluated configuration
grok-4.20-beta1
Evaluation type
human preference
Arena text preference / multi_turnhuman preference · 2026-10-15 · rating

grok-4.20-beta1

1,450.89 rating

Confidence interval: 1,441.56 – 1,460.23

4597 votes

Category
multi_turn
Evaluated configuration
grok-4.20-beta1
Evaluation type
human preference
Arena text preference / non_englishhuman preference · 2026-10-15 · rating

grok-4.20-beta1

1,433.05 rating

Confidence interval: 1,427.45 – 1,438.65

15112 votes

Category
non_english
Evaluated configuration
grok-4.20-beta1
Evaluation type
human preference
Arena text preference / overallhuman preference · 2026-10-15 · rating

grok-4.20-beta1

1,444.59 rating

Confidence interval: 1,440.06 – 1,449.13

27908 votes

Category
overall
Evaluated configuration
grok-4.20-beta1
Evaluation type
human preference
Arena text preference / polishhuman preference · 2026-10-15 · rating

grok-4.20-beta1

1,430.61 rating

Confidence interval: 1,406.76 – 1,454.45

582 votes

Category
polish
Evaluated configuration
grok-4.20-beta1
Evaluation type
human preference
Arena text preference / russianhuman preference · 2026-10-15 · rating

grok-4.20-beta1

1,449.43 rating

Confidence interval: 1,438.80 – 1,460.06

3305 votes

Category
russian
Evaluated configuration
grok-4.20-beta1
Evaluation type
human preference
Arena text preference / spanishhuman preference · 2026-10-15 · rating

grok-4.20-beta1

1,437.90 rating

Confidence interval: 1,418.48 – 1,457.33

951 votes

Category
spanish
Evaluated configuration
grok-4.20-beta1
Evaluation type
human preference

Run it locally

Inference · memory estimate

grok-4.20-beta1

Not enough verified architecture information for a reliable VRAM estimate.

Capabilities

Model license
Proprietary