U

leaderboards

Track model performance across benchmarks

🥈

G

Gemini 3.1 Pro

Google

78.2

🥇

O

GPT-5.2 Pro

OpenAI

78.9

🥉

A

Claude Opus 4.6

Anthropic

77.8

Relative

1

O

GPT-5.2 Pro

OpenAI

OpenAI78.9

—

2

G

Gemini 3.1 Pro

Google

Google78.2

−0.7

3

A

Claude Opus 4.6

Anthropic

Anthropic77.8

−1.1

4

O

GPT-5.2

OpenAI

OpenAI75.7

−3.2

5

G

Gemini 3 Pro

Google

Google74.1

−4.8

6

O

GPT-5

OpenAI

OpenAI73.0

−5.9

7

G

Gemini 3 Flash

Google

Google71.3

−7.7

8

D

DeepSeek-R2

DeepSeek

DeepSeek71.2

−7.7

9

O

o1

OpenAI

OpenAI69.8

−9.1

10

O

GPT-4.5

OpenAI

OpenAI68.0

−11.0

11

Q

Qwen3 VL 235B

Alibaba

Alibaba66.2

−12.7

12

O

o1-preview

OpenAI

OpenAI65.9

−13.0

13

Mi

Mistral Large 2

Mistral AI

Mistral AI60.3

−18.6

14

M

LLaMA 3.3 70B

Meta

Meta55.9

−23.0

15

P

Phi-4 14B

Microsoft

Microsoft51.5

−27.5

15

Models ranked

69.2

Average score

GPT-5.2 Pro

Current leader