The best LLMs for your use case:

1DeepSeek-V3.1DeepSeek

Advanced Mixture-of-Experts model with improved efficiency and performance over V3, delivering enhanced reasoning capabilities at competitive cost.

Speed:

Intelligence:

Price: (1M Tokens)

$0.71

Inputs:

ImageText

JSON Mode:

Function Calling:

Benchmarks:

#2

WebDevArena

Code

1364
#2

LiveCodeBench

Code

74.8
#1

Aider Polyglot

Code

76.3
#1

SimpleQA

General Knowledge

93.4
#2

GPQA-Diamond

General Knowledge

80.1
#2

MMLU-Pro

General Knowledge

84.8
#2

LMArena

Chat

1419
#9

LiveBench

General Knowledge

70.1
#2

WebDevArena

Code

1364
#2

LiveCodeBench

Code

74.8
#1

Aider Polyglot

Code

76.3
#1

SimpleQA

General Knowledge

93.4
#2

GPQA-Diamond

General Knowledge

80.1
#2

MMLU-Pro

General Knowledge

84.8
#2

LMArena

Chat

1419
#9

LiveBench

General Knowledge

70.1
Try it out
2DeepSeek-R1-0528DeepSeek

SOTA reasoning model trained with reinforcement learning, delivering strong performance on math, code, and logic tasks.

Speed:

Intelligence:

Price: (1M Tokens)

$3.00

Inputs:

ImageText

JSON Mode:

Function Calling:

Benchmarks:

#4

LiveCodeBench

Code

73.3
#2

Aider Polyglot

Code

71.4
#1

WebDevArena

Code

1409
#1

GPQA-Diamond

General Knowledge

81
#1

MMLU-Pro

General Knowledge

85
#1

LongBenchv2

Summarization

59.3
#3

MGSM

Multilingual

92.4
#4

LMArena

Chat

1411
#4

SimpleQA

General Knowledge

27.8
#6

EQBench

Creative Writing

1270
#8

LiveBench

General Knowledge

70.1
#17

BFCL

Agents and Function Calling

37
#4

LiveCodeBench

Code

73.3
#2

Aider Polyglot

Code

71.4
#1

WebDevArena

Code

1409
#1

GPQA-Diamond

General Knowledge

81
#1

MMLU-Pro

General Knowledge

85
#1

LongBenchv2

Summarization

59.3
#3

MGSM

Multilingual

92.4
#4

LMArena

Chat

1411
#4

SimpleQA

General Knowledge

27.8
#6

EQBench

Creative Writing

1270
#8

LiveBench

General Knowledge

70.1
#17

BFCL

Agents and Function Calling

37
Try it out

Use case:

Code