LogoLMSpeed
Back to models

Model compare

GPT-5.2 vs GPT-5.4

The readout for GPT-5.2 and GPT-5.4, before the detailed comparison sheet.

Model A

GPT-5.2

OpenAI

Contender
vs

Model B

GPT-5.4

OpenAI

Leading

Key Takeaways

Weighted outcome: GPT-5.4. Benchmark capability categories carry 80%, while price, API performance, and availability carry 20%.

Decision read

GPT-5.4

GPT-5.4 has the higher weighted result; Model A / B score 23.4 to 76.6.

Evidence depth

134 data points

Includes 35 benchmark rows, 0 audit samples, and 10 provider examples.

Selection signal

Start with GPT-5.4

The charts below split 45 high-signal samples across speed, scores, and audit health.

Change comparison

Switch either side of this report to compare another model with the same LMSpeed data pipeline.

Select a different model to open a new comparison URL.

Comparison sheet

This report only uses LMSpeed data for GPT-5.2 and GPT-5.4: pricing, speed aggregates, third-party benchmark scores, and shared provider samples.

Model compare
GPT-5.2
GPT-5.4
Overall leaderContenderLeading
Weighted overall score23.4 pts76.6 pts
Benchmark category leads1 categories6 categories
Operational advantagesCheapest input price, Average speed, First-token latencyFree providers, Provider coverage
Context window400K tokens1.1M tokens
Max output128K tokens128K tokens
Modalities

Input

FileImageText

Output

Text

Input

TextImageFile

Output

Text
Features
Text inputImage inputFile inputText outputTool callingStructured outputsJSON modeReasoning
Text inputImage inputFile inputText outputTool callingStructured outputsJSON modeReasoning

The overall result weights benchmark capability categories at 80% and price, API speed/latency, and availability at 20%. Recent test volume does not affect the winner, and missing benchmark categories are excluded.

Model metadata

Model compare
GPT-5.2
GPT-5.4
DeveloperOpenAIOpenAI
ReleasedDec 2025Mar 2026
ParametersNo dataNo data
TokenizerGPTGPT
Knowledge cutoffNo dataNo data
OpenRouter IDopenai/gpt-5.2openai/gpt-5.4
ReferencesNo dataNo data

When to choose each model

This report only uses LMSpeed data for GPT-5.2 and GPT-5.4: pricing, speed aggregates, third-party benchmark scores, and shared provider samples.

GPT-5.2

GPT-5.2 is stronger in benchmark categories (Instruction following) and operational dimensions (Cheapest input price, Average speed, First-token latency).

GPT-5.4

GPT-5.4 is stronger in benchmark categories (Agents, Coding, Reasoning, Knowledge, Math) and operational dimensions (Free providers, Provider coverage).

Benchmark score comparison

Third-party benchmark profile synced into LMSpeed; only metrics available for both models are shown.

Category performance

Compare benchmark category scores on a 0-100 scale. Select a category to inspect the gap.

Model A coverage
7 / 8
Model B coverage
7 / 8
Shared
7 shared categories

Avg. score

GPT-5.2

51.8

Avg. score

GPT-5.4

56.6

Agents

GPT-5.4 leads by 9.3

GPT-5.247
GPT-5.456.3

Coding

GPT-5.4 leads by 5.9

GPT-5.256.3
GPT-5.462.2

Reasoning

GPT-5.4 leads by 2.6

GPT-5.253.1
GPT-5.455.7

Knowledge

GPT-5.4 leads by 4.1

GPT-5.254.4
GPT-5.458.5

Math

GPT-5.4 leads by 1.5

GPT-5.255.3
GPT-5.456.8

Multilingual

No data

GPT-5.2-
GPT-5.4-

Multimodal

GPT-5.4 leads by 10.3

GPT-5.242.2
GPT-5.452.5

Instruction following

GPT-5.2 leads by 0.6

GPT-5.254.6
GPT-5.454

Professional benchmark details

Metric-level scores with benchmark source, rank depth, confidence, error, and evaluation date where available.

Aggregatereported

BenchLM overall score

SourceGPT-5.2

GPT-5.2

51.0

Rank #78/111 · confidence 4 · eval date 2025-12-11

+12.0

GPT-5.4

winner

63.0

Rank #40/111 · confidence 4 · eval date 2026-03-05

Pricingverified

Input price

SourceGPT-5.2 (Non-reasoning)

GPT-5.2

winner

$1.75/M

Rank #140/186 · confidence 4

+$0.750/M

GPT-5.4

$2.50/M

Rank #154/186 · confidence 4

Pricingverified

Blended price

SourceGPT-5.2 (Non-reasoning)

GPT-5.2

winner

$4.81/M

Rank #155/186 · confidence 4

+$0.812/M

GPT-5.4

$5.63/M

Rank #158/186 · confidence 4

Pricingverified

Output price

SourceGPT-5.2 (Non-reasoning)

GPT-5.2

winner

$14.00/M

Rank #156/186 · confidence 4

+$1.00/M

GPT-5.4

$15.00/M

Rank #159/186 · confidence 4

Reasoningreported

ARC-AGI-2

SourceGPT-5.2

GPT-5.2

52.9

Rank #15/19 · confidence 4 · eval date 2025-12-11

+21.1

GPT-5.4

winner

74.0

Rank #10/19 · confidence 4 · eval date 2026-03-05

Reasoningreported

AA-LCR

SourceGPT-5.2

GPT-5.2

winner

82.7

Rank #11/106 · confidence 4 · eval date 2025-12-11

+0.7

GPT-5.4

82.0

Rank #14/106 · confidence 4 · eval date 2026-03-05

Reasoningreported

CritPt

SourceGPT-5.2

GPT-5.2

11.6

Rank #31/88 · confidence 4 · eval date 2025-12-11

+11.8

GPT-5.4

winner

23.4

Rank #11/88 · confidence 4 · eval date 2026-03-05

Reasoningreported

BenchLM Reasoning score

SourceGPT-5.2

GPT-5.2

53.7

Rank #55/67 · confidence 4 · eval date 2025-12-11

+3.5

GPT-5.4

winner

57.2

Rank #51/67 · confidence 4 · eval date 2026-03-05

Reasoningverified

HLE

SourceGPT-5.2 (Non-reasoning)

GPT-5.2

8.0%

Rank #134/216 · confidence 4

+3.3%

GPT-5.4

winner

11.3%

Rank #116/216 · confidence 4

Reasoningverified

GPQA

SourceGPT-5.2 (Non-reasoning)

GPT-5.2

71.2%

Rank #138/218 · confidence 4

+3.6%

GPT-5.4

winner

74.8%

Rank #122/218 · confidence 4

Codingreported

Vibe Code Bench

SourceGPT-5.2

GPT-5.2

53.5

Rank #6/33 · confidence 4 · eval date 2025-12-11

+13.9

GPT-5.4

winner

67.4

Rank #3/33 · confidence 4 · eval date 2026-03-05

Codingreported

AA-SciCode

SourceGPT-5.2

GPT-5.2

52.1

Rank #28/113 · confidence 4 · eval date 2025-12-11

+4.5

GPT-5.4

winner

56.6

Rank #10/113 · confidence 4 · eval date 2026-03-05

Codingreported

SWE-bench Pro

SourceGPT-5.2

GPT-5.2

55.6

Rank #33/50 · confidence 4 · eval date 2025-12-11

+2.1

GPT-5.4

winner

57.7

Rank #21/50 · confidence 4 · eval date 2026-03-05

Codingreported

BenchLM Coding score

SourceGPT-5.2

GPT-5.2

winner

46.5

Rank #64/87 · confidence 4 · eval date 2025-12-11

+3.7

GPT-5.4

42.8

Rank #69/87 · confidence 4 · eval date 2026-03-05

Mathreported

FrontierMath v2 (Tiers 1-3)

SourceGPT-5.2

GPT-5.2

40.7

Rank #10/47 · confidence 4 · eval date 2025-12-11

+6.9

GPT-5.4

winner

47.6

Rank #7/47 · confidence 4 · eval date 2026-03-05

Mathreported

FrontierMath v2 (Tier 4)

SourceGPT-5.2

GPT-5.2

18.8

Rank #12/36 · confidence 4 · eval date 2025-12-11

+8.3

GPT-5.4

winner

27.1

Rank #9/36 · confidence 4 · eval date 2026-03-05

Mathreported

BenchLM Math score

SourceGPT-5.2

GPT-5.2

57.5

Rank #31/63 · confidence 4 · eval date 2025-12-11

+7.0

GPT-5.4

winner

64.5

Rank #24/63 · confidence 4 · eval date 2026-03-05

Knowledgereported

AA-Omniscience Hallucination Rate

SourceGPT-5.2

GPT-5.2

81.2

Rank #40/106 · confidence 4 · eval date 2025-12-11

+10.5

GPT-5.4

winner

91.7

Rank #13/106 · confidence 4 · eval date 2026-03-05

Knowledgereported

AA-HLE

SourceGPT-5.2

GPT-5.2

37.7

Rank #33/114 · confidence 4 · eval date 2025-12-11

+6.0

GPT-5.4

winner

43.7

Rank #15/114 · confidence 4 · eval date 2026-03-05

Knowledgereported

AA-Omniscience Accuracy

SourceGPT-5.2

GPT-5.2

44.3

Rank #27/106 · confidence 4 · eval date 2025-12-11

+6.5

GPT-5.4

winner

50.8

Rank #15/106 · confidence 4 · eval date 2026-03-05

Knowledgereported

AA-GPQA Diamond

SourceGPT-5.2

GPT-5.2

90.3

Rank #32/113 · confidence 4 · eval date 2025-12-11

+1.7

GPT-5.4

winner

92.0

Rank #21/113 · confidence 4 · eval date 2026-03-05

Knowledgereported

BenchLM Knowledge score

SourceGPT-5.2

GPT-5.2

winner

77.2

Rank #23/83 · confidence 4 · eval date 2025-12-11

+2.9

GPT-5.4

74.3

Rank #30/83 · confidence 4 · eval date 2026-03-05

Knowledgereported

Artificial Analysis Intelligence Index

SourceGPT-5.2

GPT-5.2

30.4

Rank #47/117 · confidence 4 · eval date 2025-12-11

+8.5

GPT-5.4

winner

39.0

Rank #28/117 · confidence 4 · eval date 2026-03-05

Multimodalreported

MMMU-Pro

SourceGPT-5.2

GPT-5.2

79.5

Rank #11/31 · confidence 4 · eval date 2025-12-11

+1.7

GPT-5.4

winner

81.2

Rank #7/31 · confidence 4 · eval date 2026-03-05

Multimodalreported

CharXiv

SourceGPT-5.2

GPT-5.2

82.1

Rank #13/30 · confidence 4 · eval date 2025-12-11

+0.7

GPT-5.4

winner

82.8

Rank #12/30 · confidence 4 · eval date 2026-03-05

Multimodalreported

BenchLM Multimodal Grounded score

SourceGPT-5.2

GPT-5.2

66.3

Rank #31/56 · confidence 4 · eval date 2025-12-11

+3.0

GPT-5.4

winner

69.3

Rank #28/56 · confidence 4 · eval date 2026-03-05

Multimodalreported

Design Arena Website

SourceGPT-5.2

GPT-5.2

1209.0

Rank #47/78 · confidence 4 · eval date 2025-12-11

+25.0

GPT-5.4

winner

1234.0

Rank #42/78 · confidence 4 · eval date 2026-03-05

Instruction followingreported

BenchLM Instruction Following score

SourceGPT-5.2

GPT-5.2

winner

92.6

Rank #10/52 · confidence 4 · eval date 2025-12-11

+2.0

GPT-5.4

90.6

Rank #16/52 · confidence 4 · eval date 2026-03-05

Instruction followingreported

AA-IFBench

SourceGPT-5.2

GPT-5.2

winner

75.4

Rank #18/84 · confidence 4 · eval date 2025-12-11

+1.5

GPT-5.4

73.9

Rank #21/84 · confidence 4 · eval date 2026-03-05

Agentsreported

τ²-bench results

SourceGPT-5.2

GPT-5.2

84.8

Rank #46/82 · confidence 4 · eval date 2025-12-11

+14.1

GPT-5.4

winner

98.9

Rank #2/82 · confidence 4 · eval date 2026-03-05

Agentsreported

Gert Labs

SourceGPT-5.2

GPT-5.2

46.5

Rank #28/49 · confidence 4 · eval date 2025-12-11

+18.4

GPT-5.4

winner

64.9

Rank #4/49 · confidence 4 · eval date 2026-03-05

Agentsreported

JobBench

SourceGPT-5.2

GPT-5.2

34.3

Rank #10/24 · confidence 4 · eval date 2025-12-11

+4.6

GPT-5.4

winner

38.9

Rank #7/24 · confidence 4 · eval date 2026-03-05

Agentsreported

OSWorld-Verified

SourceGPT-5.2

GPT-5.2

47.3

Rank #26/27 · confidence 4 · eval date 2025-12-11

+27.7

GPT-5.4

winner

75.0

Rank #11/27 · confidence 4 · eval date 2026-03-05

Agentsreported

BrowseComp

SourceGPT-5.2

GPT-5.2

65.8

Rank #25/32 · confidence 4 · eval date 2025-12-11

+16.9

GPT-5.4

winner

82.7

Rank #16/32 · confidence 4 · eval date 2026-03-05

Agentsreported

BenchLM Agentic score

SourceGPT-5.2

GPT-5.2

23.1

Rank #71/76 · confidence 4 · eval date 2025-12-11

+48.9

GPT-5.4

winner

72.0

Rank #28/76 · confidence 4 · eval date 2026-03-05

API audit comparison

Latest completed audits from shared providers, with four safety and integrity score groups plus report links.

Provider
GPT-5.2
GPT-5.4
No completed audits are available from shared providers yet.

Provider examples

Speed aggregates and input/output pricing share each provider row for real API selection and migration cost checks.

Provider
GPT-5.2
GPT-5.4
Mars HK125 tests

GPT-5.2

speed / latency

N/A / N/A

input / output

No data

GPT-5.4

speed / latency

49 tok/s / 6084ms

input / output

No data

Mars HK100 tests

GPT-5.2

speed / latency

N/A / N/A

input / output

No data

GPT-5.4

speed / latency

57 tok/s / 3556ms

input / output

No data

天宫造物50 tests

GPT-5.2

speed / latency

N/A / N/A

input / output

No data

GPT-5.4

speed / latency

50 tok/s / 7305ms

input / output

No data

QuicklyAPI35 tests

GPT-5.2

speed / latency

N/A / N/A

input / output

No data

GPT-5.4

speed / latency

50 tok/s / 2467ms

input / output

No data

星见雅 API35 tests

GPT-5.2

speed / latency

52 tok/s / 2671ms

input / output

No data

GPT-5.4

speed / latency

49 tok/s / 5435ms

input / output

No data

180txt API

GPT-5.2

gpt-5.2-2025-12-11

speed / latency

No data

input / output

$0/M/$0/M

GPT-5.4

gpt-5.4-2026-03-05

speed / latency

No data

input / output

$0/M/$0/M

APIMart

GPT-5.2

gpt-5.2

speed / latency

No data

input / output

$1.34/M/$10.74/M

GPT-5.4

gpt-5.4

speed / latency

No data

input / output

$0/M/$0/M

Dext API

GPT-5.2

gpt-5-2

speed / latency

No data

input / output

$0/M/$0/M

GPT-5.4

gpt-5-4

speed / latency

No data

input / output

$0/M/$0/M

Moyanjdc API

GPT-5.2

gpt-5.2

speed / latency

No data

input / output

$0/M/$0/M

GPT-5.4

gpt-5.4

speed / latency

No data

input / output

$0/M/$0/M

兔子API

GPT-5.2

gpt-5.2

speed / latency

No data

input / output

$0/request

GPT-5.4

gpt-5.4-2026-03-05

speed / latency

No data

input / output

$0/request

FAQ

Weighted outcome: GPT-5.4. Benchmark capability categories carry 80%, while price, API performance, and availability carry 20%.

Why is this comparison indexable?
It has 6 verifiable comparison points, and both models have pricing or benchmark data.
Are missing metrics invented?
No. Metrics without LMSpeed data are omitted from this report.

Rankings are based on community-submitted tests and periodic health probes. Advisory only, not official data.