AI Intelligence · Models

AI Model Directory

Compare 12+ AI models with benchmarks, pricing, and capabilities.Select up to 3 to compare side-by-side.

12 models·12 new this month·12 open source·12 added today
#1

Llama 3.1 40B Instruct

TODAY HOT

Meta AI · Released 2026-03-15

Open Source API

An advanced instruction-tuned model from the Llama family, offering strong performance for general language tasks and coding assistance.

TextCodeReasoning
82
MMLU
78
HumanEval
55
MATH
40
GPQA
64K tokens context40BN/A (self-hosted) in
#2

Mistral Tiny 12B Instruct

TODAY HOT

Mistral AI · Released 2026-04-01

Open Source API

A compact and efficient instruction-following model, ideal for edge deployments and resource-constrained environments.

TextCode
76
MMLU
70
HumanEval
48
MATH
35
GPQA
32K tokens context12BN/A (self-hosted) in
#3

Gemma 2.5 Base 15B

TODAY HOT

Google · Released 2026-02-20

Open Source API

The latest iteration of Google's lightweight open model family, providing strong foundational capabilities for various NLP tasks.

TextReasoning
75
MMLU
68
HumanEval
45
MATH
32
GPQA
16K tokens context15BN/A (self-hosted) in
#4

DeepSeek Coder 4.0 30B

TODAY HOT

DeepSeek AI · Released 2026-05-10

Open Source API

A highly capable code generation and understanding model, excelling in programming tasks across multiple languages.

TextCodeReasoning
79
MMLU
85
HumanEval
58
MATH
38
GPQA
64K tokens context30BN/A (self-hosted) in
#5

Qwen 3.5 Chat 40B

TODAY HOT

Alibaba Cloud · Released 2026-01-25

Open Source API

An advanced multimodal chat model from Alibaba, demonstrating strong capabilities in complex conversations and content creation.

TextCodeVisionReasoning
81
MMLU
75
HumanEval
52
MATH
37
GPQA
128K tokens context40BN/A (self-hosted) in
#6

InternLM 4.5 Base 35B

TODAY HOT

Shanghai AI Lab · Released 2026-04-20

Open Source API

A powerful foundational model developed by Shanghai AI Lab, offering robust general-purpose language understanding and generation.

TextReasoning
79
MMLU
72
HumanEval
50
MATH
36
GPQA
64K tokens context35BN/A (self-hosted) in
#7

Yi 6.0 Instruct 50B

TODAY HOT

01.AI · Released 2026-03-01

Open Source API

The latest instruction-tuned model from 01.AI, known for its extensive context window and strong performance in complex reasoning.

TextCodeReasoning
80
MMLU
76
HumanEval
53
MATH
39
GPQA
128K tokens context50BN/A (self-hosted) in
#8

Mixtral 8x7B Instruct v2

TODAY HOT

Mistral AI · Released 2026-01-10

Open Source API

An updated Mixture-of-Experts model, offering high-quality responses and efficient inference for demanding applications.

TextCodeReasoning
80
MMLU
79
HumanEval
56
MATH
41
GPQA
64K tokens context47B (effective)N/A (self-hosted) in
#9

Phi-4 7B Instruct

TODAY HOT

Microsoft · Released 2026-03-20

Open Source API

The next generation of the Phi series, delivering impressive reasoning capabilities despite its compact size.

TextReasoning
72
MMLU
65
HumanEval
42
MATH
30
GPQA
16K tokens context7BN/A (self-hosted) in
#10

Nous Hermes 3 Mistral 35B

TODAY HOT

Nous Research · Released 2026-02-15

Open Source API

A community-driven, instruction-tuned model built on Mistral architecture, optimized for superior chat and reasoning performance.

TextCodeReasoning
78
MMLU
74
HumanEval
53
MATH
37
GPQA
32K tokens context35BN/A (self-hosted) in
#11

StableLM Beta 20B

TODAY HOT

Stability AI · Released 2026-04-10

Open Source API

A mid-sized model from Stability AI's language series, providing a balanced performance for various generative tasks.

TextCode
74
MMLU
69
HumanEval
46
MATH
33
GPQA
32K tokens context20BN/A (self-hosted) in
#12

Mamba-Next 25B

TODAY HOT

Mamba Team · Released 2026-05-05

Open Source API

The next evolution of the Mamba architecture, known for its state-space model efficiency and long context processing.

TextReasoning
77
MMLU
71
HumanEval
49
MATH
34
GPQA
64K tokens context25BN/A (self-hosted) in

Want detailed model comparisons?

Visit our Benchmark Center for side-by-side comparisons across all key metrics.

View Benchmark Center