Gemini 2.5-Pro Benchmark Analysis

Gemini 2.5 Pro是谷歌发布的Gemini 2.5系列模型中最强的一个。

Benchmark Results

Gemini 2.5-Pro

Benchmark Results

综合评估

6 evaluations

Benchmark / mode

Score

Rank/total

GPQA DiamondThinking

86.40

21 / 153

MMLU ProNormal

11 / 112

LiveBenchThinking

71.92

13 / 52

ARC-AGIThinking

27 / 42

HLEThinking

21.60

50 / 105

ARC-AGI-2Thinking

4.90

25 / 34

常识问答

1 evaluations

Benchmark / mode

Score

Rank/total

SimpleQANormal

9 / 44

编程与软件工程

2 evaluations

Benchmark / mode

Score

Rank/total

LiveCodeBenchNormal

77.10

21 / 103

SWE-bench VerifiedThinking

67.20

50 / 87

数学推理

9 evaluations

Benchmark / mode

Score

Rank/total

MATH-500Normal

98.80

1 / 42

AIME 2024Normal

9 / 62

AIME2025Thinking

41 / 105

IMO-ProofBenchThinking

55.20

3 / 16

IMO 2024Thinking

2 / 10

IMO-ProofBench AdvancedThinking

17.60

4 / 8

IMO 2025Thinking

15.20

3 / 9

FrontierMathNormal

15 / 52

FrontierMath - Tier 4Normal

4.20

12 / 32

写作和创作

1 evaluations

Benchmark / mode

Score

Rank/total

Creative WritingNormal

85.85

8 / 22

AI Agent - 工具使用

2 evaluations

Benchmark / mode

Score

Rank/total

Terminal Bench 2.0Thinking + With tools

32.60

20 / 20

Terminal-BenchThinking

25.30

28 / 35

多模态理解

1 evaluations

Benchmark / mode

Score

Rank/total

MMMUThinking

5 / 17

常识推理

1 evaluations

Benchmark / mode

Score

Rank/total

Simple BenchThinking

62.40

2 / 27

Agent能力评测

3 evaluations

Benchmark / mode

Score

Rank/total

Aider-PolyglotThinking

83.10

2 / 26

τ²-Bench - TelecomThinking + With tools

26 / 29

Terminal Bench HardThinking + With tools

13 / 14

指令跟随

1 evaluations

Benchmark / mode

Score

Rank/total

IF BenchThinking + With tools

23 / 25

AI Agent - 信息收集

1 evaluations

Benchmark / mode

Score

Rank/total

BrowseCompThinking + With tools

7.80

26 / 27

生产力知识

1 evaluations

Benchmark / mode

Score

Rank/total

GDPval-AAThinking

11 / 11

长上下文能力

1 evaluations

Benchmark / mode

Score

Rank/total

AA-LCRThinking

6 / 12

与其他模型对比

Gemini 2.5-Pro Benchmark Analysis

In-depth Analysis

Benchmark Results

Benchmark Results

综合评估

常识问答

编程与软件工程

数学推理

写作和创作

AI Agent - 工具使用

多模态理解

常识推理

Agent能力评测

指令跟随

AI Agent - 信息收集

生产力知识

长上下文能力

References