Skip to main content

Head-to-head comparison

GPT-5.4 vs. GPT-5.4 mini

Compare published benchmark results from matching versions and cohorts, API pricing, and technical specifications. Any documented test setup differences remain visible.

shared result series
118
distinct benchmarks
16
latest retrieval
2026-09-04

Change model selection

Specifications

General specifications

Release date, availability, context window, architecture, and published parameter counts at a glance. Undisclosed details remain marked as unavailable.

Specifications for GPT-5.4 and GPT-5.4 mini
AttributeGPT-5.4GPT-5.4 mini
ProviderOpenAIOpenAI
ReleasedMar 2026Mar 2026
AvailabilityActiveActive
Model typeProprietaryProprietary
ParametersNot publishedNot published
ArchitectureNot publishedNot published
Context window1,050,000 tokens400,000 tokens
Knowledge cutoffAugust 31, 2025August 31, 2025
Technical sourcesModelModel

Pricing

API pricing compared

The table compares official standard API rates per 1M tokens, including published cache pricing. Batch, fast, priority, regional surcharges, tool calls, and cloud platform pricing are excluded.

API pricing for GPT-5.4 and GPT-5.4 mini
Price typeGPT-5.4GPT-5.4 mini
API input$2.5 ≤272K / $5 >272K$0.75
API output$15 ≤272K / $22.5 >272K$4.5
Cache read$0.25 ≤272K / $0.5 >272K$0.07
Cache write (5 min.)Not listedNot listed
Cache write (1 hr.)Not listedNot listed
Price verified07/30/2026 Source07/30/2026 Source
GPT-5.4GPT-5.4 mini
API inputUSD per 1M tokens, base tier
GPT-5.4$2.5
GPT-5.4 miniLower price$0.75
API outputUSD per 1M tokens, base tier
GPT-5.4$15
GPT-5.4 miniLower price$4.5
Cache readUSD per 1M tokens, base tier
GPT-5.4$0.25
GPT-5.4 miniLower price$0.075

Key to the marks:better valuebetter, but not conclusivetie

Pricing by provider

Direct developer pricing and provider endpoint tariffs routed through OpenRouter are listed separately. Prices are in USD per 1M tokens. Context-dependent direct tiers plus regional, Flex, Priority, and other OpenRouter tariffs remain individually identifiable.

Provider pricing for GPT-5.4 and GPT-5.4 mini
ModelProviderPurchase route and tariffInput per 1M tokensOutput per 1M tokensSource
GPT-5.4OpenAIDirect from developerUp to 272,000 context tokens$2.5$15Source07/30/2026
OpenAIDirect from developerAbove 272,000 context tokens$5$22.5Source07/30/2026
Amazon BedrockVia OpenRouteramazon-bedrock/us-east-1$2.75$16.5Source09/07/2026
AzureVia OpenRouterazure$2.5$15Source09/07/2026
AzureVia OpenRouterazure/eu$2.75$16.5Source09/07/2026
AzureVia OpenRouterazure/us$2.75$16.5Source09/07/2026
OpenAIVia OpenRouteropenai$2.5$15Source09/07/2026
OpenAIVia OpenRouteropenai/fast$5$30Source09/07/2026
OpenAIVia OpenRouteropenai/flexLowest listed price: $1.25Lowest listed price: $7.5Source09/07/2026
GPT-5.4 miniOpenAIDirect from developerStandard API$0.75$4.5Source07/30/2026
AzureVia OpenRouterazure$0.75$4.5Source09/07/2026
AzureVia OpenRouterazure/us$0.825$4.95Source09/07/2026
OpenAIVia OpenRouteropenai$0.75$4.5Source09/07/2026
OpenAIVia OpenRouteropenai/fast$1.5$9Source09/07/2026
OpenAIVia OpenRouteropenai/flexLowest listed price: $0.375Lowest listed price: $2.25Source09/07/2026

Performance

Shared benchmarks

Only published results with the same benchmark version, task, metric, and comparison cohort are paired. Different reasoning levels, agents, harnesses, or output limits appear directly in the table.

Capability profile from matched benchmarks

Each axis averages directly matched benchmark families. A value of 100 means the stronger value within this pair, not a universal quality score.

255075100CodingReasoningMathematicsKnowledgeFinanceLegalAgentic tasksMultimodal
GPT-5.4
GPT-5.4 mini
Coding: 100 / 70.7 (7 benchmark families)
Reasoning: 100 / 94.3 (2 benchmark families)
Mathematics: 100 / 96.7 (1 benchmark family)
Knowledge: 100 / 97.1 (2 benchmark families)
Finance: 99.6 / 94.6 (4 benchmark families)
Legal: 100 / 98.8 (2 benchmark families)
Agentic tasks: 100 / 75.8 (1 benchmark family)
Multimodal: 100 / 90.6 (1 benchmark family)

89 matched rows from 16 benchmark families were considered. The radar shows 8 of 11 comparable categories.

GPT-5.4GPT-5.4 mini
Code MigrationOverall. Higher is better. No documented setup difference
GPT-5.4Winner34.984 %
GPT-5.4 mini12.936 %
CorpFin v2Overall. Higher is better. No documented setup difference
GPT-5.4Winner65.268 %
GPT-5.4 mini60.917 %
GPQA DiamondOverall. Higher is better. No documented setup difference
GPT-5.4Winner91.666 %
GPT-5.4 mini83.08 %
Harvey's Legal Agent BenchmarkOverall · Task fully resolved. Higher is better. No documented setup difference
GPT-5.4Tie0 %
GPT-5.4 miniTie0 %
IOIOverall. Higher is better. No documented setup difference
GPT-5.4Winner67.834 %
GPT-5.4 mini6.417 %
LiveCodeBench v6Overall. Higher is better. No documented setup difference
GPT-5.4Winner84.141 %
GPT-5.4 mini81.465 %
LMArena Text, Style ControlOverall. Higher is better. Test setup not documented as identical
GPT-5.4Higher measured value1,465.527
GPT-5.4 mini1,448.118
MMLU-ProOverall. Higher is better. No documented setup difference
GPT-5.4Winner87.482 %
GPT-5.4 mini84.554 %
MMMU-ProOverall. Higher is better. No documented setup difference
GPT-5.4Winner87.514 %
GPT-5.4 mini79.249 %
MortgageTaxOverall. Higher is better. No documented setup difference
GPT-5.4Winner68.323 %
GPT-5.4 mini63.514 %
ProgramBenchOverall. Higher is better. No documented setup difference
GPT-5.4Tie0 %
GPT-5.4 miniTie0 %
SAGEOverall. Higher is better. No documented setup difference
GPT-5.443.312 %
GPT-5.4 miniWinner50.813 %

Key to the marks:better valuebetter, but not conclusivetie

Benchmark scores for GPT-5.4 and GPT-5.4 mini
Benchmark and taskGPT-5.4GPT-5.4 miniSource
Code MigrationOverall
Test details
No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is betterGPT-5.4: Snapshot: gpt-5.4-2026-03-05
GPT-5.4 mini: Snapshot: gpt-5.4-mini-2026-03-17
Winner34.984 %
12.936 %
Vals AI
Source details
Data as of: 09/03/2026
CorpFin v2Overall
Test details
No documented difference in the test setupVersion: 2Metric: accuracyScoring: Higher is betterGPT-5.4: Snapshot: gpt-5.4-2026-03-05
GPT-5.4 mini: Snapshot: gpt-5.4-mini-2026-03-17
Winner65.268 %
60.917 %
Vals AI
Source details
Data as of: 08/12/2026
GPQA DiamondOverall
Test details
No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is betterGPT-5.4: Snapshot: gpt-5.4-2026-03-05
GPT-5.4 mini: Snapshot: gpt-5.4-mini-2026-03-17
Winner91.666 %
83.08 %
Vals AI
Source details
Data as of: 09/01/2026
Harvey's Legal Agent BenchmarkOverall · Task fully resolved
Test details
No documented difference in the test setupVersion: 1Metric: task resolution rateScoring: Higher is betterGPT-5.4: Snapshot: gpt-5.4-2026-03-05
GPT-5.4 mini: Snapshot: gpt-5.4-mini-2026-03-17
Tie0 %
Tie0 %
Vals AI
Source details
Data as of: 09/03/2026
IOIOverall
Test details
No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is betterGPT-5.4: Snapshot: gpt-5.4-2026-03-05
GPT-5.4 mini: Snapshot: gpt-5.4-mini-2026-03-17
Winner67.834 %
6.417 %
Vals AI
Source details
Data as of: 08/09/2026
LiveCodeBench v6Overall
Test details
No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is betterGPT-5.4: Snapshot: gpt-5.4-2026-03-05
GPT-5.4 mini: Snapshot: gpt-5.4-mini-2026-03-17
Winner84.141 %
81.465 %
Vals AI
Source details
Data as of: 09/01/2026
LMArena Text, Style ControlOverall
Test details
Test setup not documented as identicalNo direct winnerVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.4: Snapshot: gpt-5.4, Harness: Arena text, style control
GPT-5.4 mini: Snapshot: gpt-5.4-mini-high, Reasoning: high, Harness: Arena text, style controlGPT-5.4: Style-controlled Arena rating from 63,615 anonymous pairwise votes.GPT-5.4 mini: Style-controlled Arena rating from 59,472 anonymous pairwise votes.
Higher measured value1,465.527
95% confidence interval: ±3.807The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
1,448.118
95% confidence interval: ±3.851The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.		LMArena
Source details
Data as of: 09/01/2026
MMLU-ProOverall
Test details
No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is betterGPT-5.4: Snapshot: gpt-5.4-2026-03-05
GPT-5.4 mini: Snapshot: gpt-5.4-mini-2026-03-17
Winner87.482 %
84.554 %
Vals AI
Source details
Data as of: 09/01/2026
MMMU-ProOverall
Test details
No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is betterGPT-5.4: Snapshot: gpt-5.4-2026-03-05
GPT-5.4 mini: Snapshot: gpt-5.4-mini-2026-03-17
Winner87.514 %
79.249 %
Vals AI
Source details
Data as of: 09/01/2026
MortgageTaxOverall
Test details
No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is betterGPT-5.4: Snapshot: gpt-5.4-2026-03-05
GPT-5.4 mini: Snapshot: gpt-5.4-mini-2026-03-17
Winner68.323 %
63.514 %
Vals AI
Source details
Data as of: 09/01/2026
ProgramBenchOverall
Test details
No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is betterGPT-5.4: Snapshot: gpt-5.4-2026-03-05
GPT-5.4 mini: Snapshot: gpt-5.4-mini-2026-03-17
Tie0 %
Tie0 %
Vals AI
Source details
Data as of: 09/01/2026
SAGEOverall
Test details
No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is betterGPT-5.4: Snapshot: gpt-5.4-2026-03-05
GPT-5.4 mini: Snapshot: gpt-5.4-mini-2026-03-17
43.312 %
Winner50.813 %
Vals AI
Source details
Data as of: 09/03/2026
SWE-bench VerifiedOverall
Test details
No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is betterGPT-5.4: Snapshot: gpt-5.4-2026-03-05
GPT-5.4 mini: Snapshot: gpt-5.4-mini-2026-03-17
Winner78.2 %
73 %
Vals AI
Source details
Data as of: 09/01/2026
TaxEval v2Overall
Test details
No documented difference in the test setupVersion: 2Metric: accuracyScoring: Higher is betterGPT-5.4: Snapshot: gpt-5.4-2026-03-05
GPT-5.4 mini: Snapshot: gpt-5.4-mini-2026-03-17
Winner73.958 %
71.218 %
Vals AI
Source details
Data as of: 09/01/2026
Terminal-Bench 2.0Overall
Test details
No documented difference in the test setupVersion: 2.0Metric: accuracyScoring: Higher is betterGPT-5.4: Snapshot: gpt-5.4-2026-03-05
GPT-5.4 mini: Snapshot: gpt-5.4-mini-2026-03-17
Winner58.427 %
44.944 %
Vals AI
Source details
Data as of: 06/04/2026
Vibe Code Bench v1.1Overall
Test details
No documented difference in the test setupVersion: 1.1Metric: accuracyScoring: Higher is betterGPT-5.4: Snapshot: gpt-5.4-2026-03-05, Harness: OpenHands
GPT-5.4 mini: Snapshot: gpt-5.4-mini-2026-03-17, Harness: OpenHands
Winner67.421 %
47.969 %
Vals AI
Source details
Data as of: 09/03/2026
Code MigrationCLI
Test details
No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is betterGPT-5.4: Snapshot: gpt-5.4-2026-03-05, Reasoning: xhigh, Output limit: 128,000 tokens
GPT-5.4 mini: Snapshot: gpt-5.4-mini-2026-03-17, Reasoning: xhigh, Output limit: 128,000 tokens
Winner23.788 %
8.2 %
Vals AI
Source details
Data as of: 08/09/2026
CorpFin v2Exact Pages
Test details
No documented difference in the test setupVersion: 2Metric: accuracyScoring: Higher is betterGPT-5.4: Snapshot: gpt-5.4-2026-03-05, Reasoning: xhigh, Output limit: 128,000 tokens
GPT-5.4 mini: Snapshot: gpt-5.4-mini-2026-03-17, Reasoning: xhigh, Output limit: 128,000 tokens
Winner64.103 %
61.538 %
Vals AI
Source details
Data as of: 08/06/2026
GPQA DiamondFew-Shot CoT
Test details
No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is betterGPT-5.4: Snapshot: gpt-5.4-2026-03-05, Reasoning: xhigh, Output limit: 128,000 tokens
GPT-5.4 mini: Snapshot: gpt-5.4-mini-2026-03-17, Reasoning: xhigh, Output limit: 128,000 tokens
Winner91.414 %
84.848 %
Vals AI
Source details
Data as of: 08/09/2026
IOIIOI 2024
Test details
No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is betterGPT-5.4: Snapshot: gpt-5.4-2026-03-05, Reasoning: xhigh, Output limit: 128,000 tokens
GPT-5.4 mini: Snapshot: gpt-5.4-mini-2026-03-17, Reasoning: xhigh, Output limit: 128,000 tokens
Winner67.167 %
6.667 %
Vals AI
Source details
Data as of: 08/09/2026
LiveCodeBench v6Easy
Test details
No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is betterGPT-5.4: Snapshot: gpt-5.4-2026-03-05, Reasoning: xhigh, Output limit: 128,000 tokens
GPT-5.4 mini: Snapshot: gpt-5.4-mini-2026-03-17, Reasoning: xhigh, Output limit: 128,000 tokens
Winner99.068 %
97.826 %
Vals AI
Source details
Data as of: 08/09/2026
LMArena Text, Style ControlOverall
Test details
No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.4: Snapshot: gpt-5.4-high, Reasoning: high, Harness: Arena text, style control
GPT-5.4 mini: Snapshot: gpt-5.4-mini-high, Reasoning: high, Harness: Arena text, style controlGPT-5.4: Style-controlled Arena rating from 60,657 anonymous pairwise votes.GPT-5.4 mini: Style-controlled Arena rating from 59,472 anonymous pairwise votes.
Winner1,476.488
95% confidence interval: ±3.881The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
1,448.118
95% confidence interval: ±3.851The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.		LMArena
Source details
Data as of: 09/01/2026
MMLU-ProBiology
Test details
No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is betterGPT-5.4: Snapshot: gpt-5.4-2026-03-05, Reasoning: xhigh, Output limit: 128,000 tokens
GPT-5.4 mini: Snapshot: gpt-5.4-mini-2026-03-17, Reasoning: xhigh, Output limit: 128,000 tokens
Winner92.887 %
91.353 %
Vals AI
Source details
Data as of: 08/09/2026
MortgageTaxNumerical Extraction
Test details
No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is betterGPT-5.4: Snapshot: gpt-5.4-2026-03-05, Reasoning: xhigh, Output limit: 128,000 tokens
GPT-5.4 mini: Snapshot: gpt-5.4-mini-2026-03-17, Reasoning: xhigh, Output limit: 128,000 tokens
75.994 %
Winner79.412 %
Vals AI
Source details
Data as of: 08/09/2026
ProgramBenchAlmost Resolved
Test details
No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is betterGPT-5.4: Snapshot: gpt-5.4-2026-03-05, Reasoning: xhigh, Output limit: 128,000 tokens
GPT-5.4 mini: Snapshot: gpt-5.4-mini-2026-03-17, Reasoning: xhigh, Output limit: 128,000 tokens
Winner7 %
3 %
Vals AI
Source details
Data as of: 08/09/2026
SAGEAutomata
Test details
No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is betterGPT-5.4: Snapshot: gpt-5.4-2026-03-05, Reasoning: xhigh, Output limit: 30,000 tokens
GPT-5.4 mini: Snapshot: gpt-5.4-mini-2026-03-17, Reasoning: xhigh, Output limit: 30,000 tokens
21.428 %
Winner41.27 %
Vals AI
Source details
Data as of: 08/09/2026
SWE-bench Verified<15 min fix
Test details
No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is betterGPT-5.4: Snapshot: gpt-5.4-2026-03-05, Reasoning: xhigh, Output limit: 128,000 tokens
GPT-5.4 mini: Snapshot: gpt-5.4-mini-2026-03-17, Reasoning: xhigh, Output limit: 128,000 tokens
Winner88.144 %
85.567 %
Vals AI
Source details
Data as of: 08/08/2026
TaxEval v2Correctness
Test details
No documented difference in the test setupVersion: 2Metric: accuracyScoring: Higher is betterGPT-5.4: Snapshot: gpt-5.4-2026-03-05, Reasoning: xhigh, Output limit: 128,000 tokens
GPT-5.4 mini: Snapshot: gpt-5.4-mini-2026-03-17, Reasoning: xhigh, Output limit: 128,000 tokens
Winner64.922 %
63.042 %
Vals AI
Source details
Data as of: 08/09/2026
Terminal-Bench 2.0Easy
Test details
No documented difference in the test setupVersion: 2.0Metric: accuracyScoring: Higher is betterGPT-5.4: Snapshot: gpt-5.4-2026-03-05, Reasoning: high, Output limit: 128,000 tokens
GPT-5.4 mini: Snapshot: gpt-5.4-mini-2026-03-17, Reasoning: high, Output limit: 128,000 tokens
Winner100 %
75 %
Vals AI
Source details
Data as of: 06/04/2026
Code MigrationCOBOL
Test details
No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is betterGPT-5.4: Snapshot: gpt-5.4-2026-03-05, Reasoning: xhigh, Output limit: 128,000 tokens
GPT-5.4 mini: Snapshot: gpt-5.4-mini-2026-03-17, Reasoning: xhigh, Output limit: 128,000 tokens
Winner68.571 %
27.143 %
Vals AI
Source details
Data as of: 08/09/2026
CorpFin v2Max Fitting Context
Test details
No documented difference in the test setupVersion: 2Metric: accuracyScoring: Higher is betterGPT-5.4: Snapshot: gpt-5.4-2026-03-05, Reasoning: xhigh, Output limit: 128,000 tokens
GPT-5.4 mini: Snapshot: gpt-5.4-mini-2026-03-17, Reasoning: xhigh, Output limit: 128,000 tokens
Winner64.918 %
60.956 %
Vals AI
Source details
Data as of: 08/06/2026
GPQA DiamondZero-Shot CoT
Test details
No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is betterGPT-5.4: Snapshot: gpt-5.4-2026-03-05, Reasoning: xhigh, Output limit: 128,000 tokens
GPT-5.4 mini: Snapshot: gpt-5.4-mini-2026-03-17, Reasoning: xhigh, Output limit: 128,000 tokens
Winner91.919 %
81.313 %
Vals AI
Source details
Data as of: 08/09/2026
IOIIOI 2025
Test details
No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is betterGPT-5.4: Snapshot: gpt-5.4-2026-03-05, Reasoning: xhigh, Output limit: 128,000 tokens
GPT-5.4 mini: Snapshot: gpt-5.4-mini-2026-03-17, Reasoning: xhigh, Output limit: 128,000 tokens
Winner68.5 %
6.167 %
Vals AI
Source details
Data as of: 08/09/2026
LiveCodeBench v6Hard
Test details
No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is betterGPT-5.4: Snapshot: gpt-5.4-2026-03-05, Reasoning: xhigh, Output limit: 128,000 tokens
GPT-5.4 mini: Snapshot: gpt-5.4-mini-2026-03-17, Reasoning: xhigh, Output limit: 128,000 tokens
Winner67.714 %
61.714 %
Vals AI
Source details
Data as of: 08/09/2026
LMArena Text, Style ControlBusiness, management and finance
Test details
Test setup not documented as identicalNo direct winnerVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.4: Snapshot: gpt-5.4, Harness: Arena text, style control
GPT-5.4 mini: Snapshot: gpt-5.4-mini-high, Reasoning: high, Harness: Arena text, style controlGPT-5.4: Style-controlled Arena rating from 13,060 anonymous pairwise votes.GPT-5.4 mini: Style-controlled Arena rating from 12,113 anonymous pairwise votes.
Higher measured value1,474.871
95% confidence interval: ±6.534The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
1,459.508
95% confidence interval: ±6.653The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.		LMArena
Source details
Data as of: 09/01/2026
MMLU-ProBusiness
Test details
No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is betterGPT-5.4: Snapshot: gpt-5.4-2026-03-05, Reasoning: xhigh, Output limit: 128,000 tokens
GPT-5.4 mini: Snapshot: gpt-5.4-mini-2026-03-17, Reasoning: xhigh, Output limit: 128,000 tokens
Winner91.762 %
89.734 %
Vals AI
Source details
Data as of: 08/09/2026
MortgageTaxSemantic Extraction
Test details
No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is betterGPT-5.4: Snapshot: gpt-5.4-2026-03-05, Reasoning: xhigh, Output limit: 128,000 tokens
GPT-5.4 mini: Snapshot: gpt-5.4-mini-2026-03-17, Reasoning: xhigh, Output limit: 128,000 tokens
Winner60.652 %
47.615 %
Vals AI
Source details
Data as of: 08/09/2026
ProgramBenchFully Resolved
Test details
No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is betterGPT-5.4: Snapshot: gpt-5.4-2026-03-05, Reasoning: xhigh, Output limit: 128,000 tokens
GPT-5.4 mini: Snapshot: gpt-5.4-mini-2026-03-17, Reasoning: xhigh, Output limit: 128,000 tokens
Tie0 %
Tie0 %
Vals AI
Source details
Data as of: 08/09/2026
SAGECalculus
Test details
No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is betterGPT-5.4: Snapshot: gpt-5.4-2026-03-05, Reasoning: xhigh, Output limit: 30,000 tokens
GPT-5.4 mini: Snapshot: gpt-5.4-mini-2026-03-17, Reasoning: xhigh, Output limit: 30,000 tokens
Winner70.988 %
59.259 %
Vals AI
Source details
Data as of: 08/09/2026
SWE-bench Verified>4 hours
Test details
No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is betterGPT-5.4: Snapshot: gpt-5.4-2026-03-05, Reasoning: xhigh, Output limit: 128,000 tokens
GPT-5.4 mini: Snapshot: gpt-5.4-mini-2026-03-17, Reasoning: xhigh, Output limit: 128,000 tokens
Tie0 %
Tie0 %
Vals AI
Source details
Data as of: 08/08/2026
TaxEval v2Stepwise Reasoning
Test details
No documented difference in the test setupVersion: 2Metric: accuracyScoring: Higher is betterGPT-5.4: Snapshot: gpt-5.4-2026-03-05, Reasoning: xhigh, Output limit: 128,000 tokens
GPT-5.4 mini: Snapshot: gpt-5.4-mini-2026-03-17, Reasoning: xhigh, Output limit: 128,000 tokens
Winner82.993 %
79.395 %
Vals AI
Source details
Data as of: 08/09/2026
Terminal-Bench 2.0Hard
Test details
No documented difference in the test setupVersion: 2.0Metric: accuracyScoring: Higher is betterGPT-5.4: Snapshot: gpt-5.4-2026-03-05, Reasoning: high, Output limit: 128,000 tokens
GPT-5.4 mini: Snapshot: gpt-5.4-mini-2026-03-17, Reasoning: high, Output limit: 128,000 tokens
Winner33.333 %
24.444 %
Vals AI
Source details
Data as of: 06/04/2026
Code MigrationCode Quality
Test details
No documented difference in the test setupVersion: 1Metric: mean reviewer code-quality ratingScoring: Higher is betterGPT-5.4: Snapshot: gpt-5.4-2026-03-05, Reasoning: xhigh, Output limit: 128,000 tokens
GPT-5.4 mini: Snapshot: gpt-5.4-mini-2026-03-17, Reasoning: xhigh, Output limit: 128,000 tokens
Winner3.632 / 5
3.053 / 5
Vals AI
Source details
Data as of: 08/09/2026
CorpFin v2Shared Max Context
Test details
No documented difference in the test setupVersion: 2Metric: accuracyScoring: Higher is betterGPT-5.4: Snapshot: gpt-5.4-2026-03-05, Reasoning: xhigh, Output limit: 128,000 tokens
GPT-5.4 mini: Snapshot: gpt-5.4-mini-2026-03-17, Reasoning: xhigh, Output limit: 128,000 tokens
Winner66.783 %
60.256 %
Vals AI
Source details
Data as of: 08/06/2026
LiveCodeBench v6Medium
Test details
No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is betterGPT-5.4: Snapshot: gpt-5.4-2026-03-05, Reasoning: xhigh, Output limit: 128,000 tokens
GPT-5.4 mini: Snapshot: gpt-5.4-mini-2026-03-17, Reasoning: xhigh, Output limit: 128,000 tokens
Winner85.64 %
84.856 %
Vals AI
Source details
Data as of: 08/09/2026
LMArena Text, Style ControlBusiness, management and finance
Test details
No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.4: Snapshot: gpt-5.4-high, Reasoning: high, Harness: Arena text, style control
GPT-5.4 mini: Snapshot: gpt-5.4-mini-high, Reasoning: high, Harness: Arena text, style controlGPT-5.4: Style-controlled Arena rating from 12,346 anonymous pairwise votes.GPT-5.4 mini: Style-controlled Arena rating from 12,113 anonymous pairwise votes.
Winner1,483.779
95% confidence interval: ±6.654The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
1,459.508
95% confidence interval: ±6.653The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.		LMArena
Source details
Data as of: 09/01/2026
MMLU-ProChemistry
Test details
No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is betterGPT-5.4: Snapshot: gpt-5.4-2026-03-05, Reasoning: xhigh, Output limit: 128,000 tokens
GPT-5.4 mini: Snapshot: gpt-5.4-mini-2026-03-17, Reasoning: xhigh, Output limit: 128,000 tokens
Winner91.078 %
89.134 %
Vals AI
Source details
Data as of: 08/09/2026
ProgramBenchRaw Pass Rate
Test details
No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is betterGPT-5.4: Snapshot: gpt-5.4-2026-03-05, Reasoning: xhigh, Output limit: 128,000 tokens
GPT-5.4 mini: Snapshot: gpt-5.4-mini-2026-03-17, Reasoning: xhigh, Output limit: 128,000 tokens
Winner68.596 %
54.375 %
Vals AI
Source details
Data as of: 08/09/2026
SAGELinear Algebra
Test details
No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is betterGPT-5.4: Snapshot: gpt-5.4-2026-03-05, Reasoning: xhigh, Output limit: 30,000 tokens
GPT-5.4 mini: Snapshot: gpt-5.4-mini-2026-03-17, Reasoning: xhigh, Output limit: 30,000 tokens
68.826 %
Winner71.486 %
Vals AI
Source details
Data as of: 08/09/2026
SWE-bench Verified1-4 hours
Test details
No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is betterGPT-5.4: Snapshot: gpt-5.4-2026-03-05, Reasoning: xhigh, Output limit: 128,000 tokens
GPT-5.4 mini: Snapshot: gpt-5.4-mini-2026-03-17, Reasoning: xhigh, Output limit: 128,000 tokens
Winner50 %
33.333 %
Vals AI
Source details
Data as of: 08/08/2026
Terminal-Bench 2.0Medium
Test details
No documented difference in the test setupVersion: 2.0Metric: accuracyScoring: Higher is betterGPT-5.4: Snapshot: gpt-5.4-2026-03-05, Reasoning: high, Output limit: 128,000 tokens
GPT-5.4 mini: Snapshot: gpt-5.4-mini-2026-03-17, Reasoning: high, Output limit: 128,000 tokens
Winner69.091 %
53.939 %
Vals AI
Source details
Data as of: 06/04/2026
LMArena Text, Style ControlChinese
Test details
Test setup not documented as identicalNo direct winnerVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.4: Snapshot: gpt-5.4, Harness: Arena text, style control
GPT-5.4 mini: Snapshot: gpt-5.4-mini-high, Reasoning: high, Harness: Arena text, style controlGPT-5.4: Style-controlled Arena rating from 3,684 anonymous pairwise votes.GPT-5.4 mini: Style-controlled Arena rating from 3,298 anonymous pairwise votes.
Higher measured value1,494.074
95% confidence interval: ±10.702The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
1,477.451
95% confidence interval: ±11.125The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.		LMArena
Source details
Data as of: 09/01/2026
MMLU-ProComputer Science
Test details
No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is betterGPT-5.4: Snapshot: gpt-5.4-2026-03-05, Reasoning: xhigh, Output limit: 128,000 tokens
GPT-5.4 mini: Snapshot: gpt-5.4-mini-2026-03-17, Reasoning: xhigh, Output limit: 128,000 tokens
Winner89.756 %
87.073 %
Vals AI
Source details
Data as of: 08/09/2026
SAGEPartial Differential Equations
Test details
No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is betterGPT-5.4: Snapshot: gpt-5.4-2026-03-05, Reasoning: xhigh, Output limit: 30,000 tokens
GPT-5.4 mini: Snapshot: gpt-5.4-mini-2026-03-17, Reasoning: xhigh, Output limit: 30,000 tokens
Tie39.683 %
Tie39.683 %
Vals AI
Source details
Data as of: 08/09/2026
SWE-bench Verified15 min - 1 hour
Test details
No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is betterGPT-5.4: Snapshot: gpt-5.4-2026-03-05, Reasoning: xhigh, Output limit: 128,000 tokens
GPT-5.4 mini: Snapshot: gpt-5.4-mini-2026-03-17, Reasoning: xhigh, Output limit: 128,000 tokens
Winner76.245 %
70.881 %
Vals AI
Source details
Data as of: 08/08/2026
LMArena Text, Style ControlChinese
Test details
No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.4: Snapshot: gpt-5.4-high, Reasoning: high, Harness: Arena text, style control
GPT-5.4 mini: Snapshot: gpt-5.4-mini-high, Reasoning: high, Harness: Arena text, style controlGPT-5.4: Style-controlled Arena rating from 3,482 anonymous pairwise votes.GPT-5.4 mini: Style-controlled Arena rating from 3,298 anonymous pairwise votes.
Winner1,504.962
95% confidence interval: ±10.923The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
1,477.451
95% confidence interval: ±11.125The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.		LMArena
Source details
Data as of: 09/01/2026
MMLU-ProEconomics
Test details
No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is betterGPT-5.4: Snapshot: gpt-5.4-2026-03-05, Reasoning: xhigh, Output limit: 128,000 tokens
GPT-5.4 mini: Snapshot: gpt-5.4-mini-2026-03-17, Reasoning: xhigh, Output limit: 128,000 tokens
Winner89.692 %
88.744 %
Vals AI
Source details
Data as of: 08/09/2026
LMArena Text, Style ControlCoding
Test details
Test setup not documented as identicalNo direct winnerVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.4: Snapshot: gpt-5.4, Harness: Arena text, style control
GPT-5.4 mini: Snapshot: gpt-5.4-mini-high, Reasoning: high, Harness: Arena text, style controlGPT-5.4: Style-controlled Arena rating from 17,704 anonymous pairwise votes.GPT-5.4 mini: Style-controlled Arena rating from 16,204 anonymous pairwise votes.
Higher measured value1,513.577
95% confidence interval: ±5.956The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
1,495.819
95% confidence interval: ±6.028The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.		LMArena
Source details
Data as of: 09/01/2026
MMLU-ProEngineering
Test details
No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is betterGPT-5.4: Snapshot: gpt-5.4-2026-03-05, Reasoning: xhigh, Output limit: 128,000 tokens
GPT-5.4 mini: Snapshot: gpt-5.4-mini-2026-03-17, Reasoning: xhigh, Output limit: 128,000 tokens
Winner85.243 %
82.043 %
Vals AI
Source details
Data as of: 08/09/2026
LMArena Text, Style ControlCoding
Test details
No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.4: Snapshot: gpt-5.4-high, Reasoning: high, Harness: Arena text, style control
GPT-5.4 mini: Snapshot: gpt-5.4-mini-high, Reasoning: high, Harness: Arena text, style controlGPT-5.4: Style-controlled Arena rating from 16,441 anonymous pairwise votes.GPT-5.4 mini: Style-controlled Arena rating from 16,204 anonymous pairwise votes.
Winner1,519.961
95% confidence interval: ±6.001The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
1,495.819
95% confidence interval: ±6.028The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.		LMArena
Source details
Data as of: 09/01/2026
MMLU-ProHealth
Test details
No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is betterGPT-5.4: Snapshot: gpt-5.4-2026-03-05, Reasoning: xhigh, Output limit: 128,000 tokens
GPT-5.4 mini: Snapshot: gpt-5.4-mini-2026-03-17, Reasoning: xhigh, Output limit: 128,000 tokens
Winner82.396 %
80.073 %
Vals AI
Source details
Data as of: 08/09/2026
LMArena Text, Style ControlCreative writing
Test details
Test setup not documented as identicalNo direct winnerVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.4: Snapshot: gpt-5.4, Harness: Arena text, style control
GPT-5.4 mini: Snapshot: gpt-5.4-mini-high, Reasoning: high, Harness: Arena text, style controlGPT-5.4: Style-controlled Arena rating from 10,326 anonymous pairwise votes.GPT-5.4 mini: Style-controlled Arena rating from 9,837 anonymous pairwise votes.
Higher measured value1,436.108
95% confidence interval: ±7.048The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
1,403.981
95% confidence interval: ±7.287The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.		LMArena
Source details
Data as of: 09/01/2026
MMLU-ProHistory
Test details
No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is betterGPT-5.4: Snapshot: gpt-5.4-2026-03-05, Reasoning: xhigh, Output limit: 128,000 tokens
GPT-5.4 mini: Snapshot: gpt-5.4-mini-2026-03-17, Reasoning: xhigh, Output limit: 128,000 tokens
Winner79.265 %
75.328 %
Vals AI
Source details
Data as of: 08/09/2026
LMArena Text, Style ControlCreative writing
Test details
No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.4: Snapshot: gpt-5.4-high, Reasoning: high, Harness: Arena text, style control
GPT-5.4 mini: Snapshot: gpt-5.4-mini-high, Reasoning: high, Harness: Arena text, style controlGPT-5.4: Style-controlled Arena rating from 10,108 anonymous pairwise votes.GPT-5.4 mini: Style-controlled Arena rating from 9,837 anonymous pairwise votes.
Winner1,443.763
95% confidence interval: ±7.189The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
1,403.981
95% confidence interval: ±7.287The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.		LMArena
Source details
Data as of: 09/01/2026

Key to the marks:better valuebetter, but not conclusivetie

How to read this comparison

Winning one benchmark is not an overall verdict

Your workload, budget, and required context length matter most. A coding benchmark says little about visual understanding or agent performance.

The standard price table uses direct developer rates. The provider comparison labels OpenRouter endpoints, regions, and special tariffs separately.

Every result links to its measurement source and retrieval date. Labels distinguish vendor reports, official benchmark leaderboards, and independent evaluations. Disputed or archived results are excluded.

More matchups

Related LLM comparisons

Compare GPT-5.4 and GPT-5.4 mini with other leading models using the same data and benchmark logic.

View the complete LLM comparison