The best LLMs for your use case:

1Qwen3.5 397B-A17BQwen

Qwen's native multimodal MoE model with 397B total parameters and 17B active, featuring hybrid Gated Delta Networks for strong reasoning and vision capabilities.

Speed:

Intelligence:

Price: (1M Tokens)

$0.60 / 3.60

Inputs:

ImageText

JSON Mode:

Function Calling:

Benchmarks:

#1

Multilingual MMLU

Multilingual

88.5
#1

MMLU-Pro

General Knowledge

87.8
#1

MMMU

Multimodal - Vision

85
#2

EQBench

Creative Writing

1275
#2

LongBenchv2

Summarization

63.2
#2

BFCL

Agents and Function Calling

72.9
#2

SimpleQA

General Knowledge

54.3
#3

GPQA-Diamond

General Knowledge

88.4
#3

LMArena

Chat

1447
#4

WebDevArena

Code

1189
#5

LiveCodeBench

Code

83.6
#1

Multilingual MMLU

Multilingual

88.5
#1

MMLU-Pro

General Knowledge

87.8
#1

MMMU

Multimodal - Vision

85
#2

EQBench

Creative Writing

1275
#2

LongBenchv2

Summarization

63.2
#2

BFCL

Agents and Function Calling

72.9
#2

SimpleQA

General Knowledge

54.3
#3

GPQA-Diamond

General Knowledge

88.4
#3

LMArena

Chat

1447
#4

WebDevArena

Code

1189
#5

LiveCodeBench

Code

83.6
Try it out
2GPT-OSS 120BOpenAI

OpenAI's open-source 120B parameter model with MXFP4 quantization for efficient inference.

Speed:

Intelligence:

Price: (1M Tokens)

$0.15 / 0.60

Inputs:

ImageText

JSON Mode:

Function Calling:

Benchmarks:

#2

Multilingual MMLU

Multilingual

79.3
#4

LMArena

Chat

1355
#4

EQBench

Creative Writing

1152
#5

WebDevArena

Code

1090
#7

GPQA-Diamond

General Knowledge

73.1
#8

SimpleQA

General Knowledge

16.8
#2

Multilingual MMLU

Multilingual

79.3
#4

LMArena

Chat

1355
#4

EQBench

Creative Writing

1152
#5

WebDevArena

Code

1090
#7

GPQA-Diamond

General Knowledge

73.1
#8

SimpleQA

General Knowledge

16.8
Try it out

Use case:

Multilingual

Features:

Function Calling