GPT-5.4 mini Benchmark Details

Below are GPT-5.4 mini's benchmark scores and model comparisons. In-depth analysis is being prepared.

Benchmark Results

GPT-5.4 mini

Benchmark Results

综合评估

3 evaluations

Benchmark / mode

Score

Rank/total

GPQA Diamond

Extra-High

21 / 160

HLE

Extra-High

28.20

48 / 114

HLE

Extra-HighWith tools

41.50

24 / 114

编程与软件工程

1 evaluations

Benchmark / mode

Score

Rank/total

SWE-Bench Pro - Public

Extra-HighWith tools

54.40

6 / 19

Agent能力评测

1 evaluations

Benchmark / mode

Score

Rank/total

τ²-Bench - Telecom

Extra-HighWith tools

93.40

17 / 33

AI Agent - 工具使用

3 evaluations

Benchmark / mode

Score

Rank/total

OSWorld-Verified

Extra-HighWith tools

72.10

4 / 10

Terminal Bench 2.0

Extra-HighWith tools

6 / 24

Tool Decathlon

Extra-HighWith tools

42.90

1 / 3

与其他模型对比

Competitor Comparison

Side-by-side benchmark comparison of GPT-5.4 mini against leading peer models

GPT-5.4 mini(Current model)Haiku 4.5 Gemini 3.0 Flash

No comparison data available.

Generational Comparison

Track the evolution of the GPT-5.4 mini series across generations

GPT-5.4 mini(Current model)GPT-5-mini

No comparison data available.