AI Models Explorer
January 2026 • Comprehensive Database
Search Models
Category
All Categories
Text/Language Models
Image Generation
Video Generation
Vision-Language
Audio/Speech
Embeddings
Availability
All Types
Open Source
Private/API
Company
All Companies
Modality Quick Filters
MoE Architecture
Code Specialized
Reasoning
Multimodal
0
Total Models
0
Showing
0
Open Source
0
Avg Score
Company
Model Name
Modality
Availability
Score
Pricing
Website
Important Notes
1. Scores (0-100)
Approximate estimates based on benchmark performance, user reviews, and industry assessments
Not official scores but relative performance indicators within categories
Updated regularly as new benchmarks emerge
2. Pricing
Prices change frequently and may vary by region, volume, and specific use case
Always check official websites for current pricing
API prices typically shown per 1M tokens for text models
Image/video models may use different pricing units (per image, per second, subscriptions)
3. Availability
Private:
API/subscription access only through official channels
Open Source:
Model weights are publicly available for download and self-hosting
Open source models may have specific licenses (MIT, Apache 2.0, Llama License, etc.)
Self-hosting open source models requires computational resources (GPUs) with associated costs
4. Model Categories
MoE (Mixture of Experts):
Architecture that activates only subset of parameters per inference
Multimodal:
Handles multiple input types (text, images, audio, video)
Reasoning Models:
Specialized for complex logical reasoning and problem-solving
Code Specialized:
Optimized for programming and software development tasks
5. ComfyUI & CLI Compatibility
Most open-source image and video models (Stable Diffusion, FLUX, LTX, etc.) can be run via ComfyUI
Compatible with CLI tools and frameworks like Hugging Face Diffusers, Ollama, llama.cpp
Local execution provides full control but requires technical setup and hardware
6. Data Freshness
This database represents major models as of January 2026
AI model landscape evolves rapidly with frequent releases and updates
Not exhaustive:
Hundreds of smaller, specialized, and regional models exist
Focus is on widely-used and significant models with proven performance
7. Key Model Highlights
GLM-4.7:
#1 on LiveCodeBench (84.9%), 200K context, preserved thinking feature
Kimi K2.5:
1T parameters, 256K context, Agent Swarm with 100 concurrent agents
DeepSeek-R1:
Open reasoning model competitive with o1, MIT licensed
Z-Image-Turbo:
#1 open-source on Artificial Analysis, sub-second inference
LTX-2:
First open-source synchronized audio-video generation, true 4K@50FPS
FLUX 2 Max:
Leading open-source image model with exceptional quality