geekradar
Int
Explore GeekRadar

More to explore

Good gear All topics About us Say hello
GeekRadar / The comparison desk
COMPARE.
AI models
openai

GPT-5.4

AI model Released

API mode
Chat*
Context window
1,050,000 tokens
Maximum output
128,000 tokens
Open weights
No

Capability & preference

General capability · ECI

A cross-benchmark capability estimate. Index points, not a percentage; intervals show uncertainty.

GPT-5.4 157.1* GPT-5.4 · 95% interval 155.2–159.5

Human preference · Arena

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

GPT-5.4 1452.5 GPT-5.4 · 95% interval 1448.8–1456.1 · 67,856 votes · 2026-10-28

Model performance

Radar task score: four tests, equal weight. An experimental summary of these tasks, not a universal intelligence rating.

GPT-5.4

Radar task score

Unrated

Older results · no current score

7 benchmark results for this configuration

Scientific reasoning

GPQA Diamond · 25%

Expert-written science questions; strict multiple-choice scoring. Chance accuracy is 25%.

GPT-5.4 93.3*% Approx. 95% interval: 89.8–96.8% · Older result

Novel problem solving

Mystery Game Puzzles · 25%

Best-next-move puzzles in an undisclosed game variant. A narrow reasoning test, not general intelligence.

GPT-5.4 37.0*% Approx. 95% interval: 27.5–46.5%

Advanced mathematics

FrontierMath Tiers 1-3 · 25%

Private v2 mathematics problems with Python tools. Earlier versions and Tier 4 are excluded.

GPT-5.4 78.6*% Approx. 95% interval: 73.8–83.4%

Factual answers

SimpleQA Verified · 25%

Correct factual answers under a guessing prompt, graded by a model. Not a hallucination or abstention score.

GPT-5.4 45.1*% Approx. 95% interval: 42.0–48.2%

Competition mathematics

OTIS Mock AIME 2024-2025 · Separate result

Competition problems with model-extracted answers checked against an exact key. Earlier grading procedures are excluded.

GPT-5.4 95.3*% Approx. 95% interval: 89.0–100.0% · Older result

Research mathematics

FrontierMath Tier 4 v2 · Separate result

The hardest private FrontierMath problems, corrected v2 set, with Python tools. Separate from Tiers 1-3 in the core score.

GPT-5.4 49.0*% Approx. 95% interval: 33.3–64.7%

Visual spatial reasoning

Furniture Assembly · Separate result

Assembly-error detection from photos and manuals, with zoom and Python tools. Uses model-assisted grading; not a general vision score.

GPT-5.4 37.5*% Approx. 95% interval: 25.3–49.7%

Bars show reported task scores. Whiskers are approximate 95% intervals where standard errors are available. Different reasoning budgets are not equal-cost tests. Missing results are never zero.

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

GPT-5.4

1,519.26 rating

Confidence interval: 1,509.35 – 1,529.16

4119 votes

Category
chinese
Evaluated configuration
GPT-5.4 (High)
Evaluation type
human preference

1,509.15 rating

Confidence interval: 1,499.45 – 1,518.84

4351 votes

Category
chinese
Evaluated configuration
GPT-5.4
Evaluation type
human preference
Arena text preference / codinghuman preference · 2026-10-28 · rating

GPT-5.4

1,496.50 rating

Confidence interval: 1,490.71 – 1,502.29

17603 votes

Category
coding
Evaluated configuration
GPT-5.4 (High)
Evaluation type
human preference

1,478.42 rating

Confidence interval: 1,472.71 – 1,484.14

18947 votes

Category
coding
Evaluated configuration
GPT-5.4
Evaluation type
human preference
Arena text preference / creative_writinghuman preference · 2026-10-28 · rating

GPT-5.4

1,438.64 rating

Confidence interval: 1,431.78 – 1,445.49

11111 votes

Category
creative_writing
Evaluated configuration
GPT-5.4 (High)
Evaluation type
human preference

1,427.52 rating

Confidence interval: 1,420.75 – 1,434.28

11318 votes

Category
creative_writing
Evaluated configuration
GPT-5.4
Evaluation type
human preference
Arena text preference / englishhuman preference · 2026-10-28 · rating

GPT-5.4

1,466.83 rating

Confidence interval: 1,461.93 – 1,471.73

29324 votes

Category
english
Evaluated configuration
GPT-5.4 (High)
Evaluation type
human preference

1,451.30 rating

Confidence interval: 1,446.46 – 1,456.15

30864 votes

Category
english
Evaluated configuration
GPT-5.4
Evaluation type
human preference
Arena text preference / exclude_tieshuman preference · 2026-10-28 · rating

GPT-5.4

1,473.67 rating

Confidence interval: 1,468.86 – 1,478.49

49621 votes

Category
exclude_ties
Evaluated configuration
GPT-5.4 (High)
Evaluation type
human preference

1,451.20 rating

Confidence interval: 1,446.49 – 1,455.92

52149 votes

Category
exclude_ties
Evaluated configuration
GPT-5.4
Evaluation type
human preference
Arena text preference / experthuman preference · 2026-10-28 · rating

GPT-5.4

1,506.25 rating

Confidence interval: 1,498.00 – 1,514.50

6361 votes

Category
expert
Evaluated configuration
GPT-5.4 (High)
Evaluation type
human preference

1,481.73 rating

Confidence interval: 1,473.61 – 1,489.85

6746 votes

Category
expert
Evaluated configuration
GPT-5.4
Evaluation type
human preference
Arena text preference / frenchhuman preference · 2026-10-28 · rating

GPT-5.4

1,493.57 rating

Confidence interval: 1,479.65 – 1,507.49

2386 votes

Category
french
Evaluated configuration
GPT-5.4 (High)
Evaluation type
human preference

1,471.00 rating

Confidence interval: 1,457.09 – 1,484.91

2394 votes

Category
french
Evaluated configuration
GPT-5.4
Evaluation type
human preference
Arena text preference / germanhuman preference · 2026-10-28 · rating

GPT-5.4

1,472.11 rating

Confidence interval: 1,453.57 – 1,490.64

1093 votes

Category
german
Evaluated configuration
GPT-5.4 (High)
Evaluation type
human preference

1,441.48 rating

Confidence interval: 1,421.90 – 1,461.05

999 votes

Category
german
Evaluated configuration
GPT-5.4
Evaluation type
human preference
Arena text preference / hard_promptshuman preference · 2026-10-28 · rating

GPT-5.4

1,484.99 rating

Confidence interval: 1,480.52 – 1,489.46

42125 votes

Category
hard_prompts
Evaluated configuration
GPT-5.4 (High)
Evaluation type
human preference

1,468.65 rating

Confidence interval: 1,464.24 – 1,473.06

44738 votes

Category
hard_prompts
Evaluated configuration
GPT-5.4
Evaluation type
human preference
Arena text preference / hard_prompts_englishhuman preference · 2026-10-28 · rating

GPT-5.4

1,478.74 rating

Confidence interval: 1,473.13 – 1,484.35

19715 votes

Category
hard_prompts_english
Evaluated configuration
GPT-5.4 (High)
Evaluation type
human preference

1,463.82 rating

Confidence interval: 1,458.30 – 1,469.35

21090 votes

Category
hard_prompts_english
Evaluated configuration
GPT-5.4
Evaluation type
human preference
Arena text preference / industry_business_and_management_and_financial_operationshuman preference · 2026-10-28 · rating

GPT-5.4

1,474.75 rating

Confidence interval: 1,468.38 – 1,481.12

13094 votes

Category
industry_business_and_management_and_financial_operations
Evaluated configuration
GPT-5.4 (High)
Evaluation type
human preference

1,457.23 rating

Confidence interval: 1,450.97 – 1,463.49

13842 votes

Category
industry_business_and_management_and_financial_operations
Evaluated configuration
GPT-5.4
Evaluation type
human preference
Arena text preference / industry_entertainment_and_sports_and_mediahuman preference · 2026-10-28 · rating

GPT-5.4

1,438.03 rating

Confidence interval: 1,431.70 – 1,444.36

14150 votes

Category
industry_entertainment_and_sports_and_media
Evaluated configuration
GPT-5.4 (High)
Evaluation type
human preference

1,423.81 rating

Confidence interval: 1,417.61 – 1,430.00

14792 votes

Category
industry_entertainment_and_sports_and_media
Evaluated configuration
GPT-5.4
Evaluation type
human preference
Arena text preference / industry_legal_and_governmenthuman preference · 2026-10-28 · rating

GPT-5.4

1,485.74 rating

Confidence interval: 1,476.77 – 1,494.70

5227 votes

Category
industry_legal_and_government
Evaluated configuration
GPT-5.4 (High)
Evaluation type
human preference

1,463.01 rating

Confidence interval: 1,454.12 – 1,471.89

5371 votes

Category
industry_legal_and_government
Evaluated configuration
GPT-5.4
Evaluation type
human preference
Arena text preference / industry_life_and_physical_and_social_sciencehuman preference · 2026-10-28 · rating

GPT-5.4

1,478.50 rating

Confidence interval: 1,471.80 – 1,485.21

10634 votes

Category
industry_life_and_physical_and_social_science
Evaluated configuration
GPT-5.4 (High)
Evaluation type
human preference

1,462.53 rating

Confidence interval: 1,455.91 – 1,469.16

11082 votes

Category
industry_life_and_physical_and_social_science
Evaluated configuration
GPT-5.4
Evaluation type
human preference
Arena text preference / industry_mathematicalhuman preference · 2026-10-28 · rating

GPT-5.4

1,496.76 rating

Confidence interval: 1,486.18 – 1,507.34

3547 votes

Category
industry_mathematical
Evaluated configuration
GPT-5.4 (High)
Evaluation type
human preference

1,466.58 rating

Confidence interval: 1,456.24 – 1,476.93

3738 votes

Category
industry_mathematical
Evaluated configuration
GPT-5.4
Evaluation type
human preference
Arena text preference / industry_medicine_and_healthcarehuman preference · 2026-10-28 · rating

GPT-5.4

1,457.77 rating

Confidence interval: 1,448.50 – 1,467.04

4827 votes

Category
industry_medicine_and_healthcare
Evaluated configuration
GPT-5.4 (High)
Evaluation type
human preference

1,439.48 rating

Confidence interval: 1,430.32 – 1,448.64

5026 votes

Category
industry_medicine_and_healthcare
Evaluated configuration
GPT-5.4
Evaluation type
human preference
Arena text preference / industry_software_and_it_serviceshuman preference · 2026-10-28 · rating

GPT-5.4

1,491.34 rating

Confidence interval: 1,486.21 – 1,496.47

25478 votes

Category
industry_software_and_it_services
Evaluated configuration
GPT-5.4 (High)
Evaluation type
human preference

1,474.93 rating

Confidence interval: 1,469.85 – 1,480.01

26996 votes

Category
industry_software_and_it_services
Evaluated configuration
GPT-5.4
Evaluation type
human preference
Arena text preference / industry_writing_and_literature_and_languagehuman preference · 2026-10-28 · rating

GPT-5.4

1,457.98 rating

Confidence interval: 1,452.07 – 1,463.90

16331 votes

Category
industry_writing_and_literature_and_language
Evaluated configuration
GPT-5.4 (High)
Evaluation type
human preference

1,448.69 rating

Confidence interval: 1,442.82 – 1,454.56

16721 votes

Category
industry_writing_and_literature_and_language
Evaluated configuration
GPT-5.4
Evaluation type
human preference
Arena text preference / instruction_followinghuman preference · 2026-10-28 · rating

GPT-5.4

1,469.31 rating

Confidence interval: 1,463.88 – 1,474.73

21853 votes

Category
instruction_following
Evaluated configuration
GPT-5.4 (High)
Evaluation type
human preference

1,454.94 rating

Confidence interval: 1,449.59 – 1,460.28

23376 votes

Category
instruction_following
Evaluated configuration
GPT-5.4
Evaluation type
human preference
Arena text preference / japanesehuman preference · 2026-10-28 · rating

GPT-5.4

1,483.18 rating

Confidence interval: 1,458.70 – 1,507.66

681 votes

Category
japanese
Evaluated configuration
GPT-5.4 (High)
Evaluation type
human preference

1,449.85 rating

Confidence interval: 1,426.08 – 1,473.62

685 votes

Category
japanese
Evaluated configuration
GPT-5.4
Evaluation type
human preference
Arena text preference / koreanhuman preference · 2026-10-28 · rating

GPT-5.4

1,448.40 rating

Confidence interval: 1,429.01 – 1,467.78

1114 votes

Category
korean
Evaluated configuration
GPT-5.4 (High)
Evaluation type
human preference

1,438.42 rating

Confidence interval: 1,419.52 – 1,457.32

1161 votes

Category
korean
Evaluated configuration
GPT-5.4
Evaluation type
human preference
Arena text preference / longer_queryhuman preference · 2026-10-28 · rating

GPT-5.4

1,473.18 rating

Confidence interval: 1,467.95 – 1,478.41

28430 votes

Category
longer_query
Evaluated configuration
GPT-5.4 (High)
Evaluation type
human preference

1,464.02 rating

Confidence interval: 1,458.87 – 1,469.17

29936 votes

Category
longer_query
Evaluated configuration
GPT-5.4
Evaluation type
human preference
Arena text preference / mathhuman preference · 2026-10-28 · rating

GPT-5.4

1,487.53 rating

Confidence interval: 1,476.97 – 1,498.08

3378 votes

Category
math
Evaluated configuration
GPT-5.4 (High)
Evaluation type
human preference

1,457.77 rating

Confidence interval: 1,447.48 – 1,468.06

3559 votes

Category
math
Evaluated configuration
GPT-5.4
Evaluation type
human preference
Arena text preference / multi_turnhuman preference · 2026-10-28 · rating

GPT-5.4

1,481.79 rating

Confidence interval: 1,475.22 – 1,488.37

11977 votes

Category
multi_turn
Evaluated configuration
GPT-5.4 (High)
Evaluation type
human preference

1,463.32 rating

Confidence interval: 1,456.81 – 1,469.82

12647 votes

Category
multi_turn
Evaluated configuration
GPT-5.4
Evaluation type
human preference
Arena text preference / non_englishhuman preference · 2026-10-28 · rating

GPT-5.4

1,465.12 rating

Confidence interval: 1,460.53 – 1,469.70

35422 votes

Category
non_english
Evaluated configuration
GPT-5.4 (High)
Evaluation type
human preference

1,448.04 rating

Confidence interval: 1,443.50 – 1,452.58

36990 votes

Category
non_english
Evaluated configuration
GPT-5.4
Evaluation type
human preference
Arena text preference / overallhuman preference · 2026-10-28 · rating

GPT-5.4

1,469.08 rating

Confidence interval: 1,465.35 – 1,472.81

64748 votes

Category
overall
Evaluated configuration
GPT-5.4 (High)
Evaluation type
human preference

1,452.47 rating

Confidence interval: 1,448.82 – 1,456.12

67856 votes

Category
overall
Evaluated configuration
GPT-5.4
Evaluation type
human preference
Arena text preference / polishhuman preference · 2026-10-28 · rating

GPT-5.4

1,472.41 rating

Confidence interval: 1,455.85 – 1,488.98

1320 votes

Category
polish
Evaluated configuration
GPT-5.4 (High)
Evaluation type
human preference

1,454.71 rating

Confidence interval: 1,438.14 – 1,471.28

1303 votes

Category
polish
Evaluated configuration
GPT-5.4
Evaluation type
human preference
Arena text preference / russianhuman preference · 2026-10-28 · rating

GPT-5.4

1,480.22 rating

Confidence interval: 1,472.45 – 1,488.00

7199 votes

Category
russian
Evaluated configuration
GPT-5.4 (High)
Evaluation type
human preference

1,458.27 rating

Confidence interval: 1,450.64 – 1,465.89

7666 votes

Category
russian
Evaluated configuration
GPT-5.4
Evaluation type
human preference
Arena text preference / spanishhuman preference · 2026-10-28 · rating

GPT-5.4

1,452.85 rating

Confidence interval: 1,438.76 – 1,466.94

2066 votes

Category
spanish
Evaluated configuration
GPT-5.4 (High)
Evaluation type
human preference

1,451.95 rating

Confidence interval: 1,437.79 – 1,466.11

2033 votes

Category
spanish
Evaluated configuration
GPT-5.4
Evaluation type
human preference
Arena text preference / chinesehuman preference · 2026-10-15 · rating

GPT-5.4

1,518.50 rating

Confidence interval: 1,508.60 – 1,528.39

4141 votes

Category
chinese
Evaluated configuration
gpt-5.4-high
Evaluation type
human preference

1,508.16 rating

Confidence interval: 1,498.47 – 1,517.85

4359 votes

Category
chinese
Evaluated configuration
gpt-5.4
Evaluation type
human preference
Arena text preference / codinghuman preference · 2026-10-15 · rating

GPT-5.4

1,496.44 rating

Confidence interval: 1,490.65 – 1,502.23

17600 votes

Category
coding
Evaluated configuration
gpt-5.4-high
Evaluation type
human preference

1,478.66 rating

Confidence interval: 1,472.93 – 1,484.38

18932 votes

Category
coding
Evaluated configuration
gpt-5.4
Evaluation type
human preference
Arena text preference / creative_writinghuman preference · 2026-10-15 · rating

GPT-5.4

1,438.78 rating

Confidence interval: 1,431.93 – 1,445.64

11128 votes

Category
creative_writing
Evaluated configuration
gpt-5.4-high
Evaluation type
human preference

1,427.36 rating

Confidence interval: 1,420.59 – 1,434.14

11338 votes

Category
creative_writing
Evaluated configuration
gpt-5.4
Evaluation type
human preference
Arena text preference / englishhuman preference · 2026-10-15 · rating

GPT-5.4

1,466.84 rating

Confidence interval: 1,461.93 – 1,471.74

29330 votes

Category
english
Evaluated configuration
gpt-5.4-high
Evaluation type
human preference

1,451.35 rating

Confidence interval: 1,446.51 – 1,456.20

30852 votes

Category
english
Evaluated configuration
gpt-5.4
Evaluation type
human preference
Arena text preference / exclude_tieshuman preference · 2026-10-15 · rating

GPT-5.4

1,473.57 rating

Confidence interval: 1,468.75 – 1,478.39

49653 votes

Category
exclude_ties
Evaluated configuration
gpt-5.4-high
Evaluation type
human preference

1,450.90 rating

Confidence interval: 1,446.18 – 1,455.62

52180 votes

Category
exclude_ties
Evaluated configuration
gpt-5.4
Evaluation type
human preference
Arena text preference / experthuman preference · 2026-10-15 · rating

GPT-5.4

1,506.48 rating

Confidence interval: 1,498.23 – 1,514.74

6359 votes

Category
expert
Evaluated configuration
gpt-5.4-high
Evaluation type
human preference

1,481.75 rating

Confidence interval: 1,473.62 – 1,489.89

6735 votes

Category
expert
Evaluated configuration
gpt-5.4
Evaluation type
human preference
Arena text preference / frenchhuman preference · 2026-10-15 · rating

GPT-5.4

1,492.79 rating

Confidence interval: 1,478.85 – 1,506.74

2395 votes

Category
french
Evaluated configuration
gpt-5.4-high
Evaluation type
human preference

1,471.20 rating

Confidence interval: 1,457.27 – 1,485.13

2408 votes

Category
french
Evaluated configuration
gpt-5.4
Evaluation type
human preference
Arena text preference / germanhuman preference · 2026-10-15 · rating

GPT-5.4

1,472.65 rating

Confidence interval: 1,454.06 – 1,491.23

1095 votes

Category
german
Evaluated configuration
gpt-5.4-high
Evaluation type
human preference

1,441.53 rating

Confidence interval: 1,421.91 – 1,461.15

1001 votes

Category
german
Evaluated configuration
gpt-5.4
Evaluation type
human preference
Arena text preference / hard_promptshuman preference · 2026-10-15 · rating

GPT-5.4

1,484.87 rating

Confidence interval: 1,480.40 – 1,489.34

42140 votes

Category
hard_prompts
Evaluated configuration
gpt-5.4-high
Evaluation type
human preference

1,468.43 rating

Confidence interval: 1,464.01 – 1,472.84

44742 votes

Category
hard_prompts
Evaluated configuration
gpt-5.4
Evaluation type
human preference
Arena text preference / hard_prompts_englishhuman preference · 2026-10-15 · rating

GPT-5.4

1,478.80 rating

Confidence interval: 1,473.19 – 1,484.42

19722 votes

Category
hard_prompts_english
Evaluated configuration
gpt-5.4-high
Evaluation type
human preference

1,463.87 rating

Confidence interval: 1,458.34 – 1,469.40

21080 votes

Category
hard_prompts_english
Evaluated configuration
gpt-5.4
Evaluation type
human preference
Arena text preference / industry_business_and_management_and_financial_operationshuman preference · 2026-10-15 · rating

GPT-5.4

1,474.50 rating

Confidence interval: 1,468.13 – 1,480.87

13095 votes

Category
industry_business_and_management_and_financial_operations
Evaluated configuration
gpt-5.4-high
Evaluation type
human preference

1,457.02 rating

Confidence interval: 1,450.75 – 1,463.29

13849 votes

Category
industry_business_and_management_and_financial_operations
Evaluated configuration
gpt-5.4
Evaluation type
human preference
Arena text preference / industry_entertainment_and_sports_and_mediahuman preference · 2026-10-15 · rating

GPT-5.4

1,437.80 rating

Confidence interval: 1,431.47 – 1,444.14

14154 votes

Category
industry_entertainment_and_sports_and_media
Evaluated configuration
gpt-5.4-high
Evaluation type
human preference

1,423.61 rating

Confidence interval: 1,417.41 – 1,429.81

14794 votes

Category
industry_entertainment_and_sports_and_media
Evaluated configuration
gpt-5.4
Evaluation type
human preference
Arena text preference / industry_legal_and_governmenthuman preference · 2026-10-15 · rating

GPT-5.4

1,485.71 rating

Confidence interval: 1,476.75 – 1,494.68

5239 votes

Category
industry_legal_and_government
Evaluated configuration
gpt-5.4-high
Evaluation type
human preference

1,462.85 rating

Confidence interval: 1,453.95 – 1,471.74

5378 votes

Category
industry_legal_and_government
Evaluated configuration
gpt-5.4
Evaluation type
human preference
Arena text preference / industry_life_and_physical_and_social_sciencehuman preference · 2026-10-15 · rating

GPT-5.4

1,478.08 rating

Confidence interval: 1,471.37 – 1,484.79

10648 votes

Category
industry_life_and_physical_and_social_science
Evaluated configuration
gpt-5.4-high
Evaluation type
human preference

1,461.94 rating

Confidence interval: 1,455.31 – 1,468.57

11089 votes

Category
industry_life_and_physical_and_social_science
Evaluated configuration
gpt-5.4
Evaluation type
human preference
Arena text preference / industry_mathematicalhuman preference · 2026-10-15 · rating

GPT-5.4

1,496.64 rating

Confidence interval: 1,486.05 – 1,507.23

3544 votes

Category
industry_mathematical
Evaluated configuration
gpt-5.4-high
Evaluation type
human preference

1,465.52 rating

Confidence interval: 1,455.15 – 1,475.89

3730 votes

Category
industry_mathematical
Evaluated configuration
gpt-5.4
Evaluation type
human preference
Arena text preference / industry_medicine_and_healthcarehuman preference · 2026-10-15 · rating

GPT-5.4

1,457.99 rating

Confidence interval: 1,448.72 – 1,467.26

4839 votes

Category
industry_medicine_and_healthcare
Evaluated configuration
gpt-5.4-high
Evaluation type
human preference

1,439.49 rating

Confidence interval: 1,430.32 – 1,448.65

5037 votes

Category
industry_medicine_and_healthcare
Evaluated configuration
gpt-5.4
Evaluation type
human preference
Arena text preference / industry_software_and_it_serviceshuman preference · 2026-10-15 · rating

GPT-5.4

1,491.36 rating

Confidence interval: 1,486.23 – 1,496.49

25473 votes

Category
industry_software_and_it_services
Evaluated configuration
gpt-5.4-high
Evaluation type
human preference

1,475.00 rating

Confidence interval: 1,469.92 – 1,480.09

26974 votes

Category
industry_software_and_it_services
Evaluated configuration
gpt-5.4
Evaluation type
human preference
Arena text preference / industry_writing_and_literature_and_languagehuman preference · 2026-10-15 · rating

GPT-5.4

1,457.75 rating

Confidence interval: 1,451.83 – 1,463.66

16350 votes

Category
industry_writing_and_literature_and_language
Evaluated configuration
gpt-5.4-high
Evaluation type
human preference

1,448.69 rating

Confidence interval: 1,442.82 – 1,454.56

16729 votes

Category
industry_writing_and_literature_and_language
Evaluated configuration
gpt-5.4
Evaluation type
human preference
Arena text preference / instruction_followinghuman preference · 2026-10-15 · rating

GPT-5.4

1,469.13 rating

Confidence interval: 1,463.70 – 1,474.55

21877 votes

Category
instruction_following
Evaluated configuration
gpt-5.4-high
Evaluation type
human preference

1,454.65 rating

Confidence interval: 1,449.30 – 1,460.00

23389 votes

Category
instruction_following
Evaluated configuration
gpt-5.4
Evaluation type
human preference
Arena text preference / japanesehuman preference · 2026-10-15 · rating

GPT-5.4

1,484.69 rating

Confidence interval: 1,460.03 – 1,509.36

676 votes

Category
japanese
Evaluated configuration
gpt-5.4-high
Evaluation type
human preference

1,449.81 rating

Confidence interval: 1,425.95 – 1,473.67

685 votes

Category
japanese
Evaluated configuration
gpt-5.4
Evaluation type
human preference
Arena text preference / koreanhuman preference · 2026-10-15 · rating

GPT-5.4

1,447.57 rating

Confidence interval: 1,428.12 – 1,467.03

1112 votes

Category
korean
Evaluated configuration
gpt-5.4-high
Evaluation type
human preference

1,438.04 rating

Confidence interval: 1,419.11 – 1,456.98

1163 votes

Category
korean
Evaluated configuration
gpt-5.4
Evaluation type
human preference
Arena text preference / longer_queryhuman preference · 2026-10-15 · rating

GPT-5.4

1,473.14 rating

Confidence interval: 1,467.90 – 1,478.37

28438 votes

Category
longer_query
Evaluated configuration
gpt-5.4-high
Evaluation type
human preference

1,463.83 rating

Confidence interval: 1,458.67 – 1,468.99

29951 votes

Category
longer_query
Evaluated configuration
gpt-5.4
Evaluation type
human preference
Arena text preference / mathhuman preference · 2026-10-15 · rating

GPT-5.4

1,487.38 rating

Confidence interval: 1,476.82 – 1,497.94

3379 votes

Category
math
Evaluated configuration
gpt-5.4-high
Evaluation type
human preference

1,457.48 rating

Confidence interval: 1,447.19 – 1,467.78

3559 votes

Category
math
Evaluated configuration
gpt-5.4
Evaluation type
human preference
Arena text preference / multi_turnhuman preference · 2026-10-15 · rating

GPT-5.4

1,482.11 rating

Confidence interval: 1,475.53 – 1,488.69

11988 votes

Category
multi_turn
Evaluated configuration
gpt-5.4-high
Evaluation type
human preference

1,463.03 rating

Confidence interval: 1,456.52 – 1,469.54

12657 votes

Category
multi_turn
Evaluated configuration
gpt-5.4
Evaluation type
human preference
Arena text preference / non_englishhuman preference · 2026-10-15 · rating

GPT-5.4

1,464.95 rating

Confidence interval: 1,460.36 – 1,469.54

35444 votes

Category
non_english
Evaluated configuration
gpt-5.4-high
Evaluation type
human preference

1,447.69 rating

Confidence interval: 1,443.14 – 1,452.23

37016 votes

Category
non_english
Evaluated configuration
gpt-5.4
Evaluation type
human preference
Arena text preference / overallhuman preference · 2026-10-15 · rating

GPT-5.4

1,468.97 rating

Confidence interval: 1,465.24 – 1,472.71

64776 votes

Category
overall
Evaluated configuration
gpt-5.4-high
Evaluation type
human preference

1,452.22 rating

Confidence interval: 1,448.56 – 1,455.88

67870 votes

Category
overall
Evaluated configuration
gpt-5.4
Evaluation type
human preference
Arena text preference / polishhuman preference · 2026-10-15 · rating

GPT-5.4

1,472.06 rating

Confidence interval: 1,455.46 – 1,488.66

1318 votes

Category
polish
Evaluated configuration
gpt-5.4-high
Evaluation type
human preference

1,454.28 rating

Confidence interval: 1,437.69 – 1,470.87

1303 votes

Category
polish
Evaluated configuration
gpt-5.4
Evaluation type
human preference
Arena text preference / russianhuman preference · 2026-10-15 · rating

GPT-5.4

1,480.14 rating

Confidence interval: 1,472.35 – 1,487.92

7202 votes

Category
russian
Evaluated configuration
gpt-5.4-high
Evaluation type
human preference

1,457.90 rating

Confidence interval: 1,450.27 – 1,465.53

7671 votes

Category
russian
Evaluated configuration
gpt-5.4
Evaluation type
human preference
Arena text preference / spanishhuman preference · 2026-10-15 · rating

GPT-5.4

1,453.15 rating

Confidence interval: 1,439.02 – 1,467.28

2060 votes

Category
spanish
Evaluated configuration
gpt-5.4-high
Evaluation type
human preference

1,451.94 rating

Confidence interval: 1,437.75 – 1,466.12

2034 votes

Category
spanish
Evaluated configuration
gpt-5.4
Evaluation type
human preference

Run it locally

Inference · memory estimate

GPT-5.4

Closed weights. No supported local deployment or GPU requirement is published.

API cost

USD / month

Standard text-token estimate. Excludes taxes, caching, batch discounts, reasoning and multimodal surcharges. Provider limits and regional availability may differ.

GPT-5.4

Capabilities

Context window
1,050,000 tokens
Maximum input
1,050,000 tokens*
Maximum output
128,000 tokens
Reasoning
Yes
Tool calling
Yes
Structured output
Yes*
Open weights
No
Knowledge cutoff
2025-08-31
API mode
Chat*
Image input
Yes*
PDF input
Yes*
Parallel tool calls
Yes*
Prompt caching
Yes*
System messages
Yes*
Tool choice
Yes*
Model license
Proprietary

Input modalities

  • text
  • image
  • pdf

Output modalities

  • text

System configuration

Detailed specifications

Specifications · 46 fields
cache read input token cost
2.5e-7*
Scope and conditions
native_context
provider
openai*
mode
chat*
full_native_model_id
gpt-5.4*
pricing_scope
Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.*
cache read input token cost above 272k tokens
5.0e-7*
Scope and conditions
native_context
provider
openai*
mode
chat*
full_native_model_id
gpt-5.4*
pricing_scope
Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.*
cache read input token cost above 272k tokens batches
2.5e-7*
Scope and conditions
native_context
provider
openai*
mode
chat*
full_native_model_id
gpt-5.4*
pricing_scope
Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.*
cache read input token cost above 272k tokens flex
2.5e-7*
Scope and conditions
native_context
provider
openai*
mode
chat*
full_native_model_id
gpt-5.4*
pricing_scope
Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.*
cache read input token cost batches
1.3e-7*
Scope and conditions
native_context
provider
openai*
mode
chat*
full_native_model_id
gpt-5.4*
pricing_scope
Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.*
cache read input token cost flex
1.3e-7*
Scope and conditions
native_context
provider
openai*
mode
chat*
full_native_model_id
gpt-5.4*
pricing_scope
Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.*
cache read input token cost priority
5.0e-7*
Scope and conditions
native_context
provider
openai*
mode
chat*
full_native_model_id
gpt-5.4*
pricing_scope
Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.*
default reasoning effort
none*
Scope and conditions
native_context
provider
openai*
mode
chat*
full_native_model_id
gpt-5.4*
pricing_scope
Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.*
input cost per token
2.5e-6*
Scope and conditions
native_context
provider
openai*
mode
chat*
full_native_model_id
gpt-5.4*
pricing_scope
Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.*
input cost per token above 272k tokens
5.0e-6*
Scope and conditions
native_context
provider
openai*
mode
chat*
full_native_model_id
gpt-5.4*
pricing_scope
Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.*
input cost per token above 272k tokens batches
2.5e-6*
Scope and conditions
native_context
provider
openai*
mode
chat*
full_native_model_id
gpt-5.4*
pricing_scope
Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.*
input cost per token above 272k tokens flex
2.5e-6*
Scope and conditions
native_context
provider
openai*
mode
chat*
full_native_model_id
gpt-5.4*
pricing_scope
Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.*
input cost per token batches
1.25e-6*
Scope and conditions
native_context
provider
openai*
mode
chat*
full_native_model_id
gpt-5.4*
pricing_scope
Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.*
input cost per token flex
1.25e-6*
Scope and conditions
native_context
provider
openai*
mode
chat*
full_native_model_id
gpt-5.4*
pricing_scope
Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.*
input cost per token priority
5.0e-6*
Scope and conditions
native_context
provider
openai*
mode
chat*
full_native_model_id
gpt-5.4*
pricing_scope
Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.*
litellm provider
openai*
Scope and conditions
native_context
provider
openai*
mode
chat*
full_native_model_id
gpt-5.4*
pricing_scope
Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.*
max input tokens
1,050,000*
Scope and conditions
native_context
provider
openai*
mode
chat*
full_native_model_id
gpt-5.4*
pricing_scope
Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.*
max output tokens
128,000*
Scope and conditions
native_context
provider
openai*
mode
chat*
full_native_model_id
gpt-5.4*
pricing_scope
Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.*
max tokens
128,000*
Scope and conditions
native_context
provider
openai*
mode
chat*
full_native_model_id
gpt-5.4*
pricing_scope
Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.*
mode
chat*
Scope and conditions
native_context
provider
openai*
mode
chat*
full_native_model_id
gpt-5.4*
pricing_scope
Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.*
output cost per token
1.5e-5*
Scope and conditions
native_context
provider
openai*
mode
chat*
full_native_model_id
gpt-5.4*
pricing_scope
Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.*
output cost per token above 272k tokens
2.25e-5*
Scope and conditions
native_context
provider
openai*
mode
chat*
full_native_model_id
gpt-5.4*
pricing_scope
Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.*
output cost per token above 272k tokens batches
1.125e-5*
Scope and conditions
native_context
provider
openai*
mode
chat*
full_native_model_id
gpt-5.4*
pricing_scope
Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.*
output cost per token above 272k tokens flex
1.125e-5*
Scope and conditions
native_context
provider
openai*
mode
chat*
full_native_model_id
gpt-5.4*
pricing_scope
Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.*
output cost per token batches
7.5e-6*
Scope and conditions
native_context
provider
openai*
mode
chat*
full_native_model_id
gpt-5.4*
pricing_scope
Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.*
output cost per token flex
7.5e-6*
Scope and conditions
native_context
provider
openai*
mode
chat*
full_native_model_id
gpt-5.4*
pricing_scope
Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.*
output cost per token priority
3.0e-5*
Scope and conditions
native_context
provider
openai*
mode
chat*
full_native_model_id
gpt-5.4*
pricing_scope
Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.*
regional processing uplift multiplier eu
1.1*
Scope and conditions
native_context
provider
openai*
mode
chat*
full_native_model_id
gpt-5.4*
pricing_scope
Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.*
regional processing uplift multiplier us
1.1*
Scope and conditions
native_context
provider
openai*
mode
chat*
full_native_model_id
gpt-5.4*
pricing_scope
Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.*
supported endpoints
  • /v1/chat/completions*
  • /v1/batch*
  • /v1/responses*
Scope and conditions
native_context
provider
openai*
mode
chat*
full_native_model_id
gpt-5.4*
pricing_scope
Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.*
supported modalities
  • text*
  • image*
Scope and conditions
native_context
provider
openai*
mode
chat*
full_native_model_id
gpt-5.4*
pricing_scope
Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.*
supported output modalities
  • text*
Scope and conditions
native_context
provider
openai*
mode
chat*
full_native_model_id
gpt-5.4*
pricing_scope
Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.*
supports function calling
Yes*
Scope and conditions
native_context
provider
openai*
mode
chat*
full_native_model_id
gpt-5.4*
pricing_scope
Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.*
supports minimal reasoning effort
No*
Scope and conditions
native_context
provider
openai*
mode
chat*
full_native_model_id
gpt-5.4*
pricing_scope
Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.*
supports native streaming
Yes*
Scope and conditions
native_context
provider
openai*
mode
chat*
full_native_model_id
gpt-5.4*
pricing_scope
Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.*
supports none reasoning effort
Yes*
Scope and conditions
native_context
provider
openai*
mode
chat*
full_native_model_id
gpt-5.4*
pricing_scope
Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.*
supports parallel function calling
Yes*
Scope and conditions
native_context
provider
openai*
mode
chat*
full_native_model_id
gpt-5.4*
pricing_scope
Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.*
supports pdf input
Yes*
Scope and conditions
native_context
provider
openai*
mode
chat*
full_native_model_id
gpt-5.4*
pricing_scope
Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.*
supports prompt caching
Yes*
Scope and conditions
native_context
provider
openai*
mode
chat*
full_native_model_id
gpt-5.4*
pricing_scope
Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.*
supports reasoning
Yes*
Scope and conditions
native_context
provider
openai*
mode
chat*
full_native_model_id
gpt-5.4*
pricing_scope
Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.*
supports response schema
Yes*
Scope and conditions
native_context
provider
openai*
mode
chat*
full_native_model_id
gpt-5.4*
pricing_scope
Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.*
supports system messages
Yes*
Scope and conditions
native_context
provider
openai*
mode
chat*
full_native_model_id
gpt-5.4*
pricing_scope
Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.*
supports tool choice
Yes*
Scope and conditions
native_context
provider
openai*
mode
chat*
full_native_model_id
gpt-5.4*
pricing_scope
Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.*
supports tool search
Yes*
Scope and conditions
native_context
provider
openai*
mode
chat*
full_native_model_id
gpt-5.4*
pricing_scope
Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.*
supports vision
Yes*
Scope and conditions
native_context
provider
openai*
mode
chat*
full_native_model_id
gpt-5.4*
pricing_scope
Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.*
supports xhigh reasoning effort
Yes*
Scope and conditions
native_context
provider
openai*
mode
chat*
full_native_model_id
gpt-5.4*
pricing_scope
Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.*

* Provisional data