DeepSeek R1 Distill Llama 70B vs GPT-4o mini | grube.ai

Back to Benchmarks

DeepSeek R1 Distill Llama 70B

DeepSeekvia DeepInfra

vs

GPT-4o mini

OpenAI

DeepSeek R1 Distill Llama 70B(DeepInfra) wins 10 of 11

Index Scores

16.0

+3.4

Intelligence

12.6

11.4

+11.4

Coding

0.0

53.7

+39.0

Math

14.7

Benchmarks

79.5

+14.7

MMLU Pro

64.8

40.2

GPQA

+2.4

42.6

6.1

+2.1

HLE

4.0

26.6

+3.2

LiveCode

23.4

31.2

+8.3

SciCode

22.9

93.5

+14.6

MATH-500

78.9

67.0

+55.3

AIME

11.7

Speed

42

+12

TPS

30

Shared

Winning delta

crafted by bart stefanski