A curated index of impactful AI tools and models, that emphasizes technical merit, practical utility and Prioritizing open-source.
🔺Effective AI use requires understanding capabilities, limitations, and bias mitigation strategies. 🔺
Unlock peak productivity and navigate the AI landscape with confidence. This repository is your central hub for carefully curated AI tools, models, and learning resources designed to help developers, researchers, and professionals work smarter, not harder.
Inside, discover resources to automate tasks, enhance workflows, and stay cutting-edge:
- Categorized AI solutions (Audio, Vision, LLMs, etc.), with a focus on Open Source options.
- Discover top model insights with comprehensive rankings and leverage our LLM Model Evaluation Framework (version 0.3.1) for informed decision-making. (Version 0.4 is currently in development.)
- Practical tutorials and guides.
Quickly grasp licensing and pricing models with these indicators:
- Licensing:
(Proprietary) vs. 
(Open Source) - Pricing:
(Free) | 
(Freemium) | 
(Paid)
(Documentation: CC0 License. Code/Framework Contributions: MIT License.)
This project thrives on collaboration. Here’s how you can get involved:
- 🌟 Star this repository to follow updates.
- 💡 Contribute your favorite tools, resources, or improvements (Issues & PRs welcome!).
- 📊 Help refine our AI Model Evaluation Framework.
- 🗣️ Share your experiences and use cases.
Let's build the future of intelligent workflows together!
Master AI concepts through hands-on tutorials and practical implementations.
Local tutorials
- How to run LLMs locally on your machine - Deploy and operate Large Language Models on local hardware.
- Integrating AI Models into Your Integrated Development Environment (IDE) - Configure development environments for seamless AI model integration.
- Local Image Generation with Fooocus: A Comprehensive Tutorial- Implement local image generation using open-source AI models.
- How to Use AI Privately - Implement privacy-preserving practices for AI implementation.
Online tutorials
- AI Prompt Engineering Tutor - Interactive platform for mastering prompt engineering methodologies. Developed by Loic Baconnier.
- Prompt Engineering Interactive Tutorial - Systematic guide to prompt optimization techniques. Created by Anthropic.
Beginner
| Title | Description | Platform |
|---|---|---|
| Fundamentals of Generative AI | Introduction to Generative AI and Large Language Models (LLMs). |  |
| Fundamentals of Responsible Generative AI | Using Generative AI responsibly. | |
| Introduction to Generative AI | An introduction to the capabilities, applications, and distinct characteristics of generative artificial intelligence (AI). |  |
| Introduction to Image Generation | Introduces diffusion models: a novel approach to machine learning that has generated remarkable results in image creation and manipulation. | |
| Introduction to Large Language Models | Introduction to large language models (LLMs) and the opportunities they present for natural language processing: use cases, limitations, and optimization strategies. | |
| Introduction to Responsible AI | The case for responsible AI: understanding its significance in ensuring that machine learning systems align with human values and promote social good. | |
| What are foundation models? | Discover how Foundation models are revolutionizing AI with their cutting-edge capabilities. |  |
| What are large language models (LLMs)? | Quick introduction to LLMs and their use cases. | |
| What is Conversational AI? | Basic understanding of how conversational AI works. |  |
| What is Generative AI? | Overview of foundational ideas and principles in generative AI. | |
| What is Generative AI? | Introduction to Generative AI by Understanding its Potential and Applications. | |
| What is NLP (natural language processing)? | Understand how Models understand our Language. | |
| What are vision language models (VLMs)? | Quick introduction to VLMs and their use cases. | |
Intermediate
| Title | Description | Platform |
|---|---|---|
| Evaluation of generative AI applications | Exploring and comparing different LLMs. | |
| Generative AI Explained | Concepts, applications, challenges, and opportunities in Generative AI. |  |
| Introduction to prompt engineering | Hands-on best practices for prompt engineering. | |
| Vision Language Models Explained | An overview of vision language models, their functionality, and usage. |  |
| What are AI hallucinations? | Learn why AI systems can generate nonsensical outputs by perceiving non-existent patterns or objects. | |
| What is Prompt Engineering? | A concise guide to the key concepts, considerations, and methodologies behind prompt engineering. | |
| What is prompt-tuning? | A lightweight method for fine-tuning AI foundation models on downstream tasks. | |
Advanced
| Title | Description | Platform |
|---|---|---|
| Augment your LLM Using Retrieval Augmented Generation | High-level overview of Retrieval Augmented Generation and its benefits for Generative AI (GenAI). | |
| Introduction to Quantization | An introduction to Quantization, a technique to reduce model size to improve training and inference speed. | |
| Mixture of Experts Explained | Overview of MoEs, how they’re trained, and the tradeoffs to consider. | |
| Preference Tuning LLMs with Direct Preference Optimization Methods | Exploration of three promising methods to align language models without reinforcement learning (or preference tuning). | |
| Prompt engineering techniques | Techniques that improve the outcome of your prompts. | |
| What is AI inferencing? | Introduction to the Principles and Methods of AI Inference. | |
| What is instruction tuning? | Learn how Instruction tuning enhances pre-trained LLMs by improving their ability to follow and execute instructions accurately. | |
| What is KV Cache Quantization | Understanding KV Cache Quantization to reduce memory usage for long-context text generation. | |
| What’s an LLM context window and why is it getting larger? | Understanding the Role of LLM Context Windows in AI. | |
| What is LLM orchestration | Understanding LLM orchestration and how it helps prompt, chain, manage and monitor LLMs | |
| What is Model Context Protocol (MCP) | Understanding MCP to connect LLMs to many different sources of context. | |
| What is reasoning in AI? | Understanding AI Reasoning and why it is usefull. | |
| What is retrieval-augmented generation? | Learn what is retrieval-augmented generation (RAG) and why it is usefull. | |
| What is reinforcement learning from human feedback (RLHF)? | Learn what is reinforcement learning from human feedback (RLHF) and why it is usefull. | |
| What is tool calling? | Understanding how LLMs interact with external tools. | |
AI-powered media processing toolsleverage Natural Language Processing (NLP) and computer vision algorithms to automate transcription and content summarization from audio-visual sources. These solutions streamline content analysis by generating accurate text outputs and key insights from multimedia data.
| Tool | Description | Licence | Pricing |
|---|---|---|---|
| Eightify | A powerful tool that utilizes YouTube AI technology to summarize videos quickly, providing users with key ideas in seconds. | |
 |
| Exemplary AI | A cloud-based tool that harnesses Artificial Intelligence (AI) and LLMs to offer transcription solutions. | |
|
| Riverside | An online studio that specializes in high-quality podcast and video recording and editing. | |
|
| SolidPoint | A range of tools that leverage AI technology to enhance productivity and efficiency in various tasks. One of its key features is the Summarize tool. | |
|
| Summarize.tech | An AI-powered tool that automatically generates summaries of long videos from YouTube. | |
|
| Summify | A powerful tool that efficiently condenses lengthy videos into concise and informative summaries. | |
|
| Voxweave | An innovative AI-powered tool that revolutionizes the interaction with YouTube videos by transforming them into concise summaries. | |
|
| WavoAI | An AI-powered tool that provides accurate transcriptions and insights from audio recordings. | |
|
Music generation algorithms utilize deep learning models to synthesize original compositions, enabling style-specific audio creation and adaptive soundtrack generation.
| Tool | Description | Licence | Pricing |
|---|---|---|---|
| Jukebox | A generative AI model developed by OpenAI that can create original music, including rudimentary singing, in a variety of genres and artist styles. | |
|
| Magenta | AI project developed by Google that explores the use of machine learning as a tool for creative applications, particularly in music and art. | |
|
| Mubert | A generative AI platform that allows users to create and stream original, AI-generated music and audio. | |
|
| MuseNet | An AI model developed by OpenAI that can generate original 4-minute musical compositions with up to 10 different instruments. | |
|
| Stable Audio | A generative AI system developed by Stability AI for creating high-quality audio and music. | |
|
| Suno | A cutting-edge AI-powered music generator that lets users create custom songs in various genres using text prompts. | |
|
Text-to-speech (TTS) systems employ neural networks for voice synthesis, converting text input into natural speech output. These models support voice customization parameters including timbre, prosody, and linguistic variations.
Note
The models are ranked according to their Arena Elo score (with higher scores indicating better performance) from the Artifical Analysis' Leaderboard.
| Organization | Model Name | Arena Elo | Licence | Pricing |
|---|---|---|---|---|
 |
TTS-1-HD | 1151 | |
 |
|
TTS-1 | 1137 | |
|
 |
Multilingual v2 | 1114 | |
|
|
Turbo v2.5 | 1110 | |
|
|
Flash v2.5 | 1108 | |
|
 |
Sonic English | 1106 | |
|
| Hexgrad | Kokoro-82M | 1091 | |
|
 |
T2A-01-HD | 1081 | |
|
|
Polly Generative | 1060 | |
|
|
Azure Neural | 1058 | |
|
|
Polly Long-form | 1058 | |
|
|
T2A-01-Turbo | 1042 | |
|
|
TTS Studio | 1039 | |
|
 |
Fish Speech 1.5 | 1034 | |
|
 |
Dialog | 1014 | |
|
| Zyphra | Zonos v0.1 | 1000 | |
|
|
3.0 Mini | 994 | |
|
 |
OpenVoice V2 | 972 | |
|
 |
Murf Speech Gen 2 | 972 | |
|
 |
LMNT | 971 | |
|
 |
Step TTS Mini | 959 | |
|
 |
XTTS V2 | 898 | |
|
 |
StyleTTS 2 | 889 | |
|
 |
MetaVoice V1 | 784 | |
|
| Tool | Description | Licence | Pricing |
|---|---|---|---|
| Audioread | A transformative tool that converts text into lifelike speech. | |
|
| Bark | A groundbreaking text-to-audio model developed by Suno, leveraging GPT-style models. | |
|
| Coqui | A pioneering project that focused on advancing generative voice technology. | |
|
| Eleven Labs | Industry leader proprietary tool for generating speech from text using deep learning. | |
|
| Listnr | A cutting-edge AI voice generator that seamlessly converts text into natural-sounding speech. | |
|
| MeloTTS | An open-source text-to-speech tool that uses deep learning to generate high-quality speech synthesis. | |
|
| Metavoice | A groundbreaking model that has been developed to create human-like speech with emotional nuances. | |
|
| Murf | A n innovative voice generator tool that revolutionizes the process of creating voiceovers. | |
|
| SpeechT5 | A cutting-edge model in speech synthesis and natural language processing that offers a unified approach to various speech-related tasks. | |
|
| Speechki | An advanced AI Realistic Voice Generator that offers over 1100 voices in more than 80 languages. | |
|
| Unrealspeech | A text-to-speech software that stands out for its human-like audio output, providing a superior listening experience. | |
|
| VoiceCraft | A state-of-the-art text-to-speech (TTS) model that can perform zero-shot speech editing and TTS on diverse audio data. | |
|
Speech recognition systems convert acoustic signals into text through automated speech recognition (ASR) models. These systems process audio input for text transcription and voice command interpretation.
Note
Models are ranked according to their Word Error Rate (%) (% of words transcribed incorrectly. Lower is better) from the Artifical Analysis' Leaderboard.
| Organization | Model Name | Word Error Rate (%) | Licence | Pricing |
|---|---|---|---|---|
|
Scribe | 7.7 | |
|
 |
Enhanced | 8.6 | |
|
 |
Universal-2 | 8.6 | |
|
|
Universal-1 | 8.7 | |
|
|
Chirp 2 | 9.8 | |
|
|
Whisper Large V3 | 10.3 | |
|
|
Whisper Large V2 | 10.6 | |
|
|
Transcribe | 11.2 | |
|
|
Chirp | 12.4 | |
|
|
Standard | 12.6 | |
|
 |
Nova-3 | 12.8 | |
|
|
distil-large-v3 | 13.0 | |
|
|
GPT-4o Transcribe | 13.2 | |
|
|
Nova-2 | 15.1 | |
|
|
Fish Speech to text | 19.1 | |
|
| Tool | Description | Models | Pricing |
|---|---|---|---|
| Amazon Web Services (AWS) | A fully managed service provided by Amazon Web Services (AWS) designed to facilitate the development of generative AI applications. | Amazon Transcribe | |
| AssemblyAI | A powerful speech recognition and audio intelligence platform. | Universal-1 | |
| Deepgram | A powerful accurate speech recognition with advanced AI capabilities and developer-friendly tools. | Nova-2 and Whisper Large V2 | |
| DeepInfra | A platform that provides scalable and cost-effective infrastructure for deploying machine learning models. | Whisper Large V3 and distil-large-v3 | |
| Fal.ai | A powerful cloud platform designed for deploying and integrating AI models into applications. | Whisper Large V3 | |
| Gladia | An advanced AI platform that specializes in real-time transcription, translation, and audio intelligence. | Whisper Large V2 | |
| A powerful service offered by Google Cloud that utilizes advanced machine learning techniques to convert spoken language into written text. | Chirp | |
|
| Groq | Specializes in high-performance AI inference with custom LPU (Language Processing Unit) hardware, offering models like Meta's Llama 3. | Whisper Large V3 and distil-large-v3 | |
| Microsoft Azure | A comprehensive suite of AI services and tools designed to help developers and organizations build, deploy, and manage AI applications at scale. | Whisper Large V2 | |
| OpenAI | A state-of-the-art automatic speech recognition (ASR) system developed by OpenAI. | Whisper Large V2 | |
| Replicate | A cloud platform that allows developers to easily run and deploy open-source machine learning models. | All Whisper Familly | |
| Rev AI | A sophisticated speech recognition platform that provides automatic speech-to-text transcription services. | Rev AI | |
| Speechmatics | A powerful AI-driven speech recognition and transcription platform. | Universal-1 | |
These systems combine multiple AI technologies to create interactive voice experiences.
| Organization | Model Familly | Best Model | Licence | Pricing |
|---|---|---|---|---|
 |
Moshi | Moshi v0.1 | |
|
| Tool | Description | Models | Pricing |
|---|---|---|---|
| OpenAI | Premium voice interface for GPT-4, offering natural conversations with high-quality voice synthesis and recognition. Features multiple voice options and seamless integration with ChatGPT. | GPT4-o | |
| Gemini | Google's conversational AI assistant offering natural voice interactions through the Gemini app. Features multilingual support, voice input/output, and integration with Google services. | Gemini 1.5 Pro | |
AI agents are autonomous software systems that execute predefined tasks through decision-making algorithms and environment interaction protocols. These systems implement adaptive learning mechanisms and inter-agent communication frameworks to achieve specified objectives.
| Tool | Description | Licence | Pricing |
|---|---|---|---|
| AgentGPT | A generative artificial intelligence tool that allows users to create autonomous AI agents capable of performing various tasks autonomously. | |
|
| Cognosys | An AI assistant that can help you automate tasks, organize your work, and perform research. | |
|
| Evo.ninja | a generalist agent that can flow between multiple agent personas to solve any task. | |
|
| Godmode | A web platform that provides access to innovative AI agents like autoGPT and babyAGI, allowing users to harness the power of autonomous AI agents. | |
|
| GPT-Engineer | An open-source AI-powered application builder that generates codebases from natural language project descriptions. | |
|
| Super AGI | An open-source autonomous AI agent framework that enables developers to build, manage, and run useful autonomous agents efficiently and reliably. | |
|
Execute predefined task sequences through algorithmic workflows to optimize process efficiency and minimize operational variance.
| Tool | Description | Licence | Pricing |
|---|---|---|---|
| Bardeen | An AI-powered automation platform that enables users to automate repetitive tasks across various applications without writing code. It offers pre-built integrations with popular tools and allows users to create custom workflows. | |
|
| Cykel | an AI company focused on developing intelligent automation solutions that can understand natural language and interact with various software and websites to automate complex digital tasks for businesses. | |
|
| Gumloop | AI-native workflow automation platform that allows users to build complex automations by visually connecting modular components on a canvas | |
|
| Lindy | An advanced automation platform designed to create custom AI assistants that streamline various business workflows without requiring coding skills. | |
|
| N8N | A free and open-source fair-code licensed workflow automation tool. It allows users to create workflows using a visual editor and connect various services to automate tasks. N8N can be self-hosted, providing users with more control over their data. | |
|
| ProFlow | an AI-powered workflow automation and optimization platform that helps businesses streamline their sales, marketing, and operations processes. | |
|
| Taskade | An all-in-one collaboration platform that combines project management, task tracking, and team communication features. It offers real-time syncing, customizable templates, and integrations with popular tools. Taskade also has AI-powered features like smart due dates and natural language processing for better task management. | |
|
| Zapier | A popular web-based automation platform that connects various apps and services to automate workflows. It offers a wide range of pre-built integrations and allows users to create custom automation rules called "Zaps" without needing to write code. Zapier's AI capabilities include filtering, formatting, and transforming data between apps. | |
|
Computer Vision (CV) frameworks implement neural architectures for visual data processing, analysis, and synthesis across image and video domains.
Caution
Use AI-generated images responsibly: Always disclose that they were generated by AI. Be mindful of intellectual property rights.
Tip
Learn prompt engineering techniques for image generation models to enhance output quality and artistic control. Follow @nickfloats on 𝕏 for valuable insights on crafting prompts that achieve your desired visual outputs.
| Tool | Description | Licence | Pricing |
|---|---|---|---|
| BRIA AI | An AI-powered model to automatically remove backgrounds from images. | |
|
| Clarity AI | AI Image Upscaler & Enhancer - free and open-source Magnific Alternative | |
|
| ImageFX | An AI-powered tool for applying various image effects and filters. | |
|
| Lensa | An AI-powered mobile app for editing and enhancing photos, particularly for portrait editing. | |
|
| Luminar Neo | An AI-powered photo editing software developed by Skylum. | |
|
| Magnific AI | an AI-powered image upscaler and enhancer designed for professionals and enthusiasts in photography, graphic design, digital art, and illustration. | |
|
| Pixlr | An AI-powered online photo editing tool. | |
|
| Removebg | An online tool that allows users to automatically remove backgrounds from images. | |
|
| ZMO AI | Comprehensive online platform offering AI-powered image editing tools. Features include background removal, object erasure, image enhancement, and creative modifications. | |
|
Note
The models are ranked according to their Elo scores (with higher scores indicating better performance) from the artificialanalysis.ai text to Image Arena and Imgsys.org Ranking. Please note that Elo scores are subject to change based on user votes and will be updated regularly to reflect the latest rankings.
To provide a comprehensive overview of the generative image model landscape, only pre-trained versions of the listed models are included in this ranking.
Due to the continuous evolution and vast number of possible fine-tuned configurations, it is impractical to comprehensively list every variant here.
| Organization | Model | Elo score | Licence | Pricing |
|---|---|---|---|---|
|
GPT-4o | 1144 | |
|
 |
Recraft V3 | 1105 | |
|
| HiDream | HiDream-I1-Dev | 1103 | |
|
| Reve AI | Reve Image 1.0 | 1098 | |
|
|
Imagen 3 | 1095 | |
|
 |
Flux1.1 Pro | 1079 | |
|
|
Flux.1 Pro | 1064 | |
|
|
MiniMax Image-01 | 1049 | |
|
 |
Midjourney v6.1 | 1045 | |
|
|
Flux.1 Dev | 1042 | |
|
 |
Ideogram v2 | 1041 | |
|
|
Midjourney v7 Alpha | 1039 | |
|
|
Midjourney v6 | 1038 | |
|
|
Ideogram v2 Turbo | 1033 | |
|
 |
Photon | 1033 | |
|
 |
Stable Diffusion 3.5 Large Turbo | 1030 | |
|
|
Stable Diffusion 3.5 Large | 1026 | |
|
 |
Infinity 8B | 1021 | |
|
|
Ideogram v1 | 1021 | |
|
|
Stable Diffusion 3 Large | 1014 | |
|
|
Flux.1 schnell | 1000 | |
|
 |
Playground v3 (beta) | 997 | |
|
|
Recraft 20B | 976 | |
|
|
Photon Flash | 996 | |
|
|
Playground v2.5 | 954 | |
|
 |
Lumina Image v2 | 950 | |
|
 |
Firefly Image 3 | 942 | |
|
|
DALLE 3 HD | 941 | |
|
|
Stable Diffusion 3.5 medium | 932 | |
|
|
DALLE 3 | 926 | |
|
|
Stable Diffusion 3 Medium | 902 | |
|
|
Stable Diffusion 3 Large Turbo | 897 | |
|
|
Stable Diffusion 1.6 | 885 | |
|
|
Stable Diffusion XL base 1.0 | 849 | |
|
|
DALLE 2 | 714 | |
|
|
Stable Diffusion 2.1 | 712 | |
|
|
Stable Diffusion 1.5 | 625 | |
|
| Tool | Description | Licence | Pricing |
|---|---|---|---|
| Craiyon | An AI-powered platform for generating artistic images and animations. | |
|
| Dall-E | An AI model developed by OpenAI that generates images from textual descriptions. | |
|
| Fal.ai | Fal.ai is a cutting-edge generative media platform designed for developers to build advanced AI applications. | |
|
| Firefly | A creative AI tool for generating images, animations, and other visual content. | |
|
| Ideogram | An advanced text-to-image generator that creates high-quality images based on text prompts. | |
|
| Krea | An advanced AI-powered platform designed for generating and enhancing visual content, including images and videos. | |
|
| Lexica | An AI art platform that generates images from textual descriptions. | |
|
| Leonardo | An open-source AI model for generating images from textual descriptions. | |
|
| Midjourney | A world-famous AI platform that generates images and visual content based on user input. | |
|
| Nightcafe | An open-source AI art platform that generates images from textual descriptions using deep learning models. | |
|
| Picasso | An AI-powered platform for generating images and animations, developed by NVIDIA. | |
|
| Removebg | An online tool that allows users to automatically remove backgrounds from images. | |
|
| Stable diffusion | An open-source AI model for generating images from textual descriptions using diffusion-based generative models. | |
|
Tip
Generate images locally using Fooocus - Deploy open-source image generation models on your hardware with our Local Image Generation with Fooocus: A Comprehensive Tutorial.
| Tool | Description | OS | Models |
|---|---|---|---|
| ComfyUI | A powerful and modular graphical user interface (GUI) for Stable Diffusion, provide users with precise control over image generation workflows. | All | All Stable Diffusion Models + Flux.1 |
| Diffusion Bee | A free, offline AI art generation tool designed specifically for macOS users. | MacOS/IOS | All Stable Diffusion Models. |
| Draw Things | A free AI-assisted image generation app available for iOS devices, including iPhones and iPads. | MacOS/IOS | All Stable Diffusion Models. |
| Fooocus | An open-source AI image generation tool designed to simplify the process of creating images using Stable Diffusion technology. | All | Stable Diffusion XL models. |
| Invoke | A leading creative engine for Stable Diffusion models. | All | All Stable Diffusion Models. |
| Stable Diffusion web UI by Automatic1111 | a popular graphical user interface (GUI) for interacting with the Stable Diffusion models. | All | All Stable Diffusion Models. |
Note
Video generation technology remains primarily concentrated among major AI research organizations, with models like OpenAI's Sora and Runway's Gen3 leading development. Current publicly available implementations are limited due to the computational complexity and proprietary nature of these systems.
This section will be updated as more open-source and accessible video generation models emerge.
Image-to-video models employ temporal diffusion algorithms to synthesize video sequences from static image inputs, generating coherent motion patterns and frame transitions.
| Organization | Model Familly | Best Model | Licence | Pricing |
|---|---|---|---|---|
 |
CogVideo | CogVideoX-5B-I2V | |
|
|
Stable Video Diffusion (SVD) | img2vid-xt | |
|
|
Stable Video Diffusion (SV3D) | sv3d | |
|
|
Stable Video Diffusion (SV4D) | sv4d | |
|
| Lightricks | LTXV | LTX-Video | |
|
 |
Wan | Wan2.1-I2V-14B-720P | |
|
Text-to-video models convert natural language descriptions into video sequences through multi-modal generation frameworks, synthesizing temporal and spatial elements from textual inputs.
Note
The models are ranked according to their Elo scores (with higher scores indicating better performance) from the artificialanalysis.ai Video Generation Arena. Please note that Elo scores are subject to change based on user votes and will be updated regularly to reflect the latest rankings.
| Organization | Best Model | Elo score | Licence | Pricing |
|---|---|---|---|---|
|
Veo 2 | 1122 | |
|
 |
Kling 1.5 (Pro) | 1053 | |
|
|
Sora | 1049 | |
|
 |
Pika 2.0 | 1039 | |
|
 |
MiniMax T2V-01 | 1039 | |
|
|
Kling 1.6 (Pro) | 1021 | |
|
|
Wan2.1-T2V-14B | 1018 | |
|
 |
HunyuanVideo | 1002 | |
|
 |
Mochi-1 | 1000 | |
|
 |
Gen-3 Alpha | 989 | |
|
|
Kling 1.0 | 969 | |
|
|
Ray 1 | 969 | |
|
|
Ray 2 | 954 | |
|
| Haiper | Haiper 2.0 | 947 | |
|
|
Pika 1.5 | 943 | |
|
|
CogVideoX-5B | 784 | |
|
| Yang Jin | Pyramid-flow | 760 | |
|
| Tool | Description | Licence | Pricing |
|---|---|---|---|
| Dream Machine | A groundbreaking text-to-video AI tool that enables users to generate high-quality, realistic video clips from simple text prompts in just minutes. | |
|
| Elai | A video creation platform that enables users to produce videos by inputting text that is then narrated by AI-generated avatars. | |
|
| Heygen | An innovative video platform that harnesses the power of generative AI to streamline the video creation process. | |
|
| Higgsfield | A pioneering foundational model company that specializes in democratizing social media content creation through AI-powered video generation and editing tools. | |
|
| Kling | An advanced video generation model developed by Kuaishou Technology, known for its capabilities in creating high-quality videos from text prompts. | |
|
| Krea | An advanced AI-powered platform designed for generating and enhancing visual content, including images and videos. | |
|
| Runway | An AI-powered platform for creatives to use machine learning models in their workflows. | |
|
| Sora | An AI model developed by OpenAI for generating videos from textual descriptions. | |
|
| Synthesia | A synthetic media generation AI tool to create AI-generated video content efficiently. | |
|
| Veo | A generative video model developed by Google, capable of producing high-quality 1080p videos. | |
|
| Vlogger | A method for text and audio-driven talking human video generation from a single input image of a person. | |
|
| Wombo | An AI-powered mobile app for creating lip-syncing videos and other creative content. | |
|
Transform text descriptions and images into detailed 3D models using AI. These Models enable rapid prototyping, asset creation, and visualization by converting natural language or visual inputs into three-dimensional objects.
| Organization | Model | Licence | Pricing |
|---|---|---|---|
|
Hunyuan3D-2 | |
|
|
InstantMesh | |
|
|
Stable-zero123 | |
|
|
TripoSR | |
|
|
stable-fast-3d | |
|
| craftsman3d | CraftsMan-v1-5 | |
|
| Ashawkey | LGM | |
|
| Jade choghari | vfusion3d | |
|
| Zhaoxi Chen | 3DTopia-XL | |
|
Data Analysis frameworks implement machine learning models for processing structured and unstructured datasets, enabling pattern recognition and statistical inference across diverse data formats.
Caution
Exercise caution with fully automated analysis results, as errors and biases may occur. Use AI tools as a complement to human judgment for more reliable insights.
| Tool | Description | Licence | Pricing |
|---|---|---|---|
| AskCSV | An AI-powered tool that allows users to ask questions about CSV data files in natural language and receive answers. | |
|
| DataSquirrel | An AI-powered data extraction and analysis tool. | |
|
| Grapha AI | An AI-powered platform for automating data analysis and generating insights. | |
|
| Hal9 | Data analytics leveraging generative AI to get insights from databases. | |
|
| Julius | An AI-powered tool for automating data entry and document processing tasks. | |
|
| Monitr | An AI-powered data extraction and analysis tool. | |
|
| Pi Exchange | A platform for building and deploying AI models. | |
|
| Research Studio | An AI-powered research assistant that helps users find, analyze, and summarize information. | |
|
| Rows AI | An AI-powered spreadsheet tool that helps users automate data analysis and manipulation tasks. | |
|
| Vizly | A tool for creating interactive data visualizations. | |
|
Enhance spreadsheet functionality in Excel and Google Sheets through AI-assisted formula generation and optimization. 
| Tool | Description | Licence | Pricing |
|---|---|---|---|
| Formulabot | A virtual assistant designed to streamline the process of creating Excel formulas by understanding natural language instructions. | |
|
| GPTExcel | An AI tool designed to generate and explain Microsoft Excel and Google Sheets formulas efficiently. | |
|
| Numerous | An AI-powered tool designed to enhance productivity and automate tasks in spreadsheet applications like Google Sheets and Microsoft Excel. | |
|
| Sheety | A tool designed to streamline the process of creating Google Sheets formulas using artificial intelligence. | |
|
Large Language Models (LLMs) are sophisticated artificial intelligence systems that have been trained on enormous amounts of text data to grasp and produce human-like language via pattern analysis.
This overview focuses on instruction-tuned models to maintain clarity. Fine-tuned variants, while numerous and continuously evolving, are not included in this comparison. See Fine-tuned Models section below for specialized implementations.
Models are ranked using our structured scoring system that balances multiple criteria, such as Entity & Publisher benchmarks, human preference, and technical features.
Note that a higher score doesn't guarantee better performance in all tasks or domains; it offers an overall assessment within the model's own size family.
Tip
Follow our tutorial to deploy LLMs locally using Ollama and Page Assist for secure, privacy-focused language processing.
Large-scale models (70+ billion parameters) : These require significant amounts of both RAM and GPU memory, often rendering local installation infeasible for most users. Consequently, such models are predominantly deployed on cloud-based platforms designed to provide the essential computational resources needed.
| Organization | Model Name | Model Sizes | Score (v0.3.1) | Context Window | Reasoning Model | Geographic Origin |
|---|---|---|---|---|---|---|
 |
DeepSeek-R1 | 685B | 79.61 | 128K | ✔️ |  |
 |
R1-1776 | 685B | 79.51 | 128K | ✔️ |  |
|
Llama-3_1-Nemotron-Ultra-253B-v1 | 253B | 76.20 | 128K | ✔️ | |
|
DeepSeek-V3 | 685B | 72.79 | 128K | ❌ | |
 |
Cogito-v1-preview-llama-70B | 70B | 68.81 | 128K | ✔️ | |
|
Hunyuan-Large | 389B | 67.98 | 128K | ❌ | |
|
MiniMax-Text-01 | 456B | 67.73 | 4M | ❌ | |
|
Qwen2.5-72B-Instruct | 72B | 67.38 | 131K | ❌ | |
 |
Llama-4-Maverick-17B-128E-Instruct | 402B | 67.14 | 1M | ❌ | |
|
Llama-3.3-70B-Instruct | 70B | 66.64 | 128K | ❌ | |
|
DeepSeek-R1-Distill-Llama-70B | 70.6B | 65.40 | 128K | ✔️ | |
 |
Command A | 111B | 64.25 | 256k | ❌ | |
 |
Athene-V2-Chat | 70B | 63.79 | 131K | ❌ | |
|
Llama-3.1-405B-Instruct | 405B | 63.49 | 128K | ❌ | |
|
Llama-4-Scout-17B-16E-Instruct | 109B | 63.29 | 10M | ❌ | |
 |
Ling-plus | 293B | 62.90 | 64k | ❌ | |
 |
Mistral Large 2 | 123B | 60.38 | 128K | ❌ |  |
 |
Jamba 1.6 Large | 399B | 55.39 | 256K | ❌ |  |
|
Mixtral-8x22B-Instruct-v0.1 | 141B | 52.97 | 65k | ❌ | |
 |
Dbrx-instruct | 132B | 52.10 | 33k | ❌ | |
|
Command R+ | 104B | 49.01 | 128k | ❌ | |
Mid-sized models (14+ billion parameters) : These models are well-suited for local deployment on high-end workstations. However, such deployments require a significant hardware investment, including a powerful GPU (24-32 GB of VRAM) and associated components, typically resulting in total costs exceeding $3,000 (or equivalent).
| Organization | Model Name | Model Sizes | Score (v0.3.1) | Context Window | Reasoning Model | Geographic Origin |
|---|---|---|---|---|---|---|
|
Llama-3_3-Nemotron-Super-49B-v1 | 49B | 69.71 | 128K | ✔️ | |
|
GLM-Z1-32B-0414 | 32.3B | 69.51 | 132K | ✔️ | |
|
Cogito-v1-preview-qwen-32B | 32.3B | 69.36 | 132K | ✔️ | |
|
GLM-4-32B-0414 | 32.3B | 68.89 | 132K | ❌ | |
 |
Reka Flash 3 | 21B | 67.16 | 128K | ✔️ | |
 |
EXAONE-Deep-32B | 32B | 66.11 | 32k | ✔️ |  |
|
QwQ-32B | 32B | 64.27 | 131K | ✔️ | |
 |
OpenThinker-32B | 32B | 63.75 | 132K | ✔️ | |
|
Qwen2.5-32B-Instruct | 32B | 62.70 | 132K | ❌ | |
|
DeepSeek-R1-Distill-Qwen-32B | 32B | 62.62 | 132K | ✔️ | |
|
Gemma-3-27b-it | 27B | 62.03 | 128k | ❌ | |
|
EXAONE-3.5-32B-Instruct | 32B | 60.94 | 32k | ❌ | |
|
Mistral-Small-3 | 23.9B | 57.84 | 128K | ❌ | |
 |
OLMo-2-0325-32B-Instruct | 32B | 57.40 | 32k | ❌ | |
|
Ling-lite | 16.8B | 55.15 | 64k | ❌ | |
|
Command R | 32.3B | 39.53 | 128K | ❌ | |
Small models (7B+ parameters) : These are lightweight and easily deployable on medium machines, offering broader accessibility. They typically require a mid-range consumer configuration, including a GPU (8-16 GB of VRAM) and associated components, with costs generally between $1,000 to $2,000 (or equivalent).
| Organization | Model Name | Model Sizes | Score (v0.3.1) | Context Window | Reasoning Model | Geographic Origin |
|---|---|---|---|---|---|---|
|
Cogito-v1-preview-qwen-14B | 14B | 66.91 | 132K | ✔️ | |
|
Llama-3.1-Nemotron-Nano-8B-v1 | 8B | 62.81 | 128K | ✔️ | |
|
DeepSeek-R1-Distill-Qwen-14B | 14B | 60.91 | 132K | ✔️ | |
|
Gemma-3-12b-it | 12B | 59.60 | 128k | ❌ | |
|
EXAONE-Deep-7.8B | 7.8B | 59.29 | 32k | ✔️ | |
|
Qwen2.5-14B-Instruct | 14B | 59.52 | 132K | ❌ | |
|
Internlm3-8b-instruct | 8B | 58.90 | 300K | ❌ | |
|
EXAONE-3.5-7.8B-Instruct | 7.8B | 58.00 | 32k | ❌ | |
|
GLM-Z1-9B-0414 | 9B | 57.51 | 132K | ✔️ | |
|
Phi-4 | 14B | 57.05 | 16k | ❌ | |
|
Cogito-v1-preview-llama-8B | 8B | 56.24 | 128K | ✔️ | |
|
Qwen2.5-7B-Instruct | 7B | 55.85 | 132K | ❌ | |
|
Granite-3.3-8b-instruct | 8B | 54.91 | 128k | ❌ | |
|
DeepSeek-R1-Distill-Qwen-7B | 7B | 53.59 | 128K | ✔️ | |
 |
Falcon3-10B-Instruct | 10B | 51.88 | 32k | ❌ |  |
|
Ministral-8B-Instruct | 8B | 51.52 | 128K | ❌ | |
|
OLMo-2-1124-13B-Instruct | 13B | 51.22 | 4k | ❌ | |
|
DeepSeek-R1-Distill-Llama-8B | 8B | 50.71 | 128K | ✔️ | |
|
Llama-3.1-8B-Instruct | 8B | 49.87 | 128K | ❌ | |
|
OLMo-2-1124-7B-Instruct | 7B | 47.49 | 4k | ❌ | |
|
GLM-4-9B-0414 | 9B | Pending | 132K | ❌ | |
Tiny models (under 7B parameters) : Designed for broad compatibility, these models run effectively on older or less powerful machines, making them accessible to a wider range of users. They typically require only 6-8 GB of RAM and can be deployed across a wide range of standard consumer hardware setups.
| Organization | Model Name | Model Sizes | Score (v0.3.1) | Context Window | Reasoning Model | Geographic Origin |
|---|---|---|---|---|---|---|
|
Cogito-v1-preview-llama-3B | 3B | 56.74 | 128K | ✔️ | |
|
Llama-3.2-3B-Instruct | 3B | 53.89 | 128K | ❌ | |
|
EXAONE-3.5-2.4B-Instruct | 2.4B | 51.69 | 32k | ❌ | |
|
EXAONE-Deep-2.4B | 2.4B | 50.90 | 32k | ✔️ | |
|
Gemma-3-4b-it | 4B | 50.20 | 128k | ❌ | |
 |
MiniCPM3-4B | 4B | 49.99 | 32K | ❌ | |
|
Granite-3.3-2b-instruct | 2B | 46.72 | 128k | ❌ | |
|
Ministral-3B-Instruct | 3B | 46.60 | 128K | ❌ | |
|
Qwen2.5-3B-Instruct | 3B | 45.95 | 32K | ❌ | |
|
Qwen2.5-1.5B-Instruct | 1.5B | 39.79 | 32K | ❌ | |
|
DeepSeek-R1-Distill-Qwen-1.5B | 1.5B | 36.96 | 128K | ✔️ | |
|
Llama-3.2-1B-Instruct | 1B | 36.32 | 128K | ❌ | |
|
Gemma-3-1b-it | 1B | 31.47 | 32k | ❌ | |
|
Qwen2.5-0.5B-Instruct | 0.5B | 31.37 | 32K | ❌ | |
Note that scores reflect an overall assessment and do not guarantee consistently superior performance in every situation.
| Organization | Model Name | Score (v0.3.1) | Context Window | Reasoning Model | Geographic Origin | Pricing |
|---|---|---|---|---|---|---|
|
Gemini 2.5 Pro | 83.22 | 1M | ✔️ | |
|
|
o3 | 79.34 | 256k | ✔️ | |
|
|
Gemini 2.5 Flash | 77.09 | 1M | ✔️ | |
|
|
o4-mini | 75.87 | 256k | ✔️ | |
|
 |
Grok-3 | 75.4 | 1M | ✔️ | |
|
|
GPT-4.1 | 72.97 | 1M | ❌ | |
|
 |
Claude 3.7 Sonnet | 72.92 | 200k | ✔️ | |
|
 |
Doubao 1.5 Pro | 71.60 | 256K | ❌ | |
|
|
GPT-4.1 mini | 70.97 | 1M | ❌ | |
|
|
Gemini 2.0 Flash | 69.99 | 1M | ❌ | |
|
 |
Kimi-k1.5 | 69.93 | Unknown | ❌ | |
|
|
Qwen 2.5 Max | 68.33 | 32K | ❌ | |
|
|
Sonar Pro | 66.87 | 200k | ❌ | |
|
 |
GLM-4-Plus | 67.12 | 1M | ❌ | |
|
|
GPT-4o | 66.32 | 128K | ❌ | |
|
|
Nova Pro | 66.20 | 300K | ❌ | |
|
|
Claude 3.5 Sonnet | 65.97 | 200k | ❌ | |
|
|
Gemini 2.0 Flash-Lite | 65.91 | 1M | ❌ | |
|
|
Step-2-16k-exp | 64.69 | 16K | ❌ | |
|
|
GPT-4.1 nano | 61.99 | 1M | ❌ | |
|
|
Reka Core | 60.18 | 128k | ❌ | |
|
|
Claude 3.5 Haiku | 59.04 | 200k | ❌ | |
|
|
GLM-4-Air | Pending | 1M | ❌ | |
|
|
GLM-4-Flash | Pending | 1M | ❌ | |
|
 |
ERNIE-4.5 | Pending | Unknown | ❌ | |
|
|
ERNIE-X1 | Pending | Unknown | ✔️ | |
|
Fine-tuned Large Language Models (LLMs) refer to AI models that have been specifically adapted for a particular domain, task, or dataset. This adaptation significantly enhances their performance and accuracy within that specific context, compared to training them on more general-purpose datasets.
Models optimized for Astrophysics and Astronomy research through specialized training datasets.
| Organization | Base Model | Finetuned Model | Model Sizes | Context Window | Knowledge Cutoff | Licence |
|---|---|---|---|---|---|---|
| AstroMLab | Llama-3.1-8B | AstroSage-8B | 8B | 128K | 2023-12 | |
| Tijmen de Haan | Llama-3.1-8B | Cosmosage-v3.1 | 8B | 128K | 2023-12 | |
| AstroMLab | Llama-2-70b-hf | Astrollama-2-70b-base_aic | 8B | 128K | 2022-09 | |
AI models specifically trained on code to assist with tasks like writing, completing, and understanding software.
Note
Models are ranked by BigCodeBench Hard set with Pass@1 metrics, measuring single-attempt problem-solving accuracy.
Generalist models can match or exceed domain-specific coding models in certain tasks. Compare Pass@1 scores across both model categories in our comparative analysis.
| Organization | Model Familly | Best Model | Model Sizes | Pass@1 | Context Window |
|---|---|---|---|---|---|
|
Qwen2.5 Coder | Qwen2.5-Coder-32B-Instruct | 32B | 30.8 | 132K |
|
Deepseek-coder | DeepSeek-Coder-V2-Instruct | 236B | 29.4 | 128K |
|
Codestral | Codestral-22B-v0.1 | 22B | 20.6 | 33K |
 |
DeepCoder | DeepCoder-14B-Preview | 14B | 20.3 | 64K |
|
CodeQwen | CodeQwen1.5-7B | 7B | 17.2 | 64K |
|
CodeGeex | Codegeex4-all-9b | 9B | 17.2 | 128K |
 |
OpenCodeInterpreter | OpenCodeInterpreter-DS-33B | 33B | 15.2 | 8K |
|
Granite Code | Granite-34b-code-instruct-8k | 34B | 14.8 | 8k |
 |
Yi-Coder | Yi-Coder-9B-Chat | 9B | 14.6 | 128K |
|
Mamba-Codestral | Mamba-Codestral-7B-v0.1 | 7B | 13.9 | 256k |
 |
Starcoder | Starcoder2-15b | 15B | 13.6 | 4K |
|
CodeLlama | CodeLlama-70b-Instruct-hf | 70B | 13.5 | 16K |
|
Codegemma | Codegemma-1.1-7b-it | 7B | 10.4 | 8K |
 |
Qwen2.5 Coder | Openhands-lm-32b-v0.1 | 32B | Pending | 132K |
|
Ling | Ling-Coder-lite | 16.8B | Pending | 8K |
Tip
Follow our integration guide to configure cloud-based and local LLM providers within your development environment.
Models optimized for Function calling tasks through specialized training.
Function calling enables LLMs to interact with external systems and tools through structured interfaces.
Note
Models are ranked by Berkeley Function Calling Leaderboard V3 Score, designed to evaluate the function calling capabilities of LLMs. it provides a comprehensive evaluation of LLMs' function calling capabilities, offering insights into their performance, cost-effectiveness, and error patterns in real-world scenarios.
| Organization | Base Model | Finetuned Model | Model Sizes | BCFL Score | Context Window | Knowledge Cutoff | Licence |
|---|---|---|---|---|---|---|---|
| Meetkai | Llama-3.1-70B-Instruct | Functionary-medium-v3.1 | 70B | 62.53 | 128K | 2023-12 | |
| Katanemo | Qwen2.5-7B | Arch-Function-7B | 7B | 59.62 | 131K | 2024-04 | |
| Team-ACE | Llama-3.1-8B-Instruct | ToolACE-8B | 8B | 58.31 | 128K | 2023-12 | |
| Salesforce | Mixtral-8x22B-Instruct-v0.1 | xLAM-8x22b-r | 141B | 58.03 | 64K | 2023-?? | |
| MadeAgents | Qwen2.5-Coder-7B-Instruct | Hammer2.0-7b | 7B | 55.19 | 131K | 2024-04 | |
| Fireworks | Llama-3.1-70B-Instruct | llama-3-firefunction-v2 | 70B | 53.12 | 128K | 2023-12 | |
|
Granite-20b-code-instruct-8k | Granite-20b-functioncalling | 20B | 49.19 | 8k | Unknown | |
|
CodeLlama-13b-Instruct-hf | NexusRaven-V2-13B | 13B | 36.98 | 8K | 2023-?? | |
Models optimized for mathematical reasoning and computation through specialized training architectures.
Note
Model rankings utilize combined performance metrics from GSM8K and MATH benchmarks, averaging scores across both frameworks to provide comprehensive evaluation standards.
| Organization | Model Familly | Best Model | Model Sizes | Score | Context Window | Licence |
|---|---|---|---|---|---|---|
|
AceMath | AceMath-72B-Instruct | 72B | 91.25 | 132K | |
|
Qwen2.5 Math | Qwen2.5-Math-72B-Instruct | 72B | 90.35 | 132K | |
 |
NuminaMath | NuminaMath-72B-CoT | 72B | 78.75 | 128K | |
|
InternLM2-Math-Plus | Internlm2-math-plus-mixtral8x22b | 141B | 74.95 | 65K | |
|
Deepseek-math | Deepseek-math-7b-instruct | 7B | 69.95 | 4K | |
|
Mathstral | Mathstral-7B-v0.1 | 7B | 66.85 | 4K | |
Other Open source Models Family Variants
| Organization | Model Familly | Best Model | Model Sizes | Score | Context Window | Licence |
|---|---|---|---|---|---|---|
|
Qwen2.5 Math | Qwen2.5-Math-7B-Instruct | 7B | 89.9 | 132K | |
| Qwen2.5-Math-1.5B-Instruct | 1.5B | 81.8 | 132K | |
||
|
InternLM2-Math-Plus | Internlm2-math-plus-20B | 20B | 70.75 | 4K | |
| Internlm2-math-plus-7B | 7B | 69.4 | 4K | |
||
| Internlm2-math-plus-1.8B | 1.8B | 47.9 | 4K | |
||
|
NuminaMath | NuminaMath-7B-CoT | 7B | 65.3 | 4K | |
Models optimized for Role Play tasks through specialized training datasets.
Role-playing in LLMs is a technique where the model assumes a specific character, profession, or persona to generate more focused and contextually relevant responses.
| Organization | Base Model | Finetuned Model | Model Sizes | Context Window | Knowledge Cutoff | Licence |
|---|---|---|---|---|---|---|
| Steelskull | Llama-3.3-70B-Instruct | L3.3-MS-Nevoria-70b | 70B | 128K | 2023-12 | |
| BosonAI | Llama-3-70B-Instruct | Higgs-Llama-3-70B | 70B | 32K | 2023-04 | |
| ResplendentAI | Mistral-Small-Instruct-2409 | Pantheon-RP-Pure-1.6.2-22b-Small | 22B | 128K | 2023-12 | |
| Latitude Games | Mistral-Nemo-Instruct-2407 | Wayfarer-12B | 13B | 128K | 2023-12 | |
| Oxygen | Qwen2.5-14B-Instruct | Oxy-1-small | 14.8B | 131K | 2024-04 | |
Models modified to operate without standard content filtering mechanisms, enabling unrestricted response generation beyond typical LLM safeguards.
| Organization | Best Model | Model Sizes | Context Window | Reasoning Model | Geographic Origin |
|---|---|---|---|---|---|
 |
Hermes-3-Llama-3.1-405B | 405B | 128K | ❌ | |
| Maxime Labonne | Llama-3.1-70B-Instruct-lorablated | 70B | 128K | ❌ | |
 |
Dolphin3.0-Llama3.1-8B | 8B | 128K | ❌ | |
| Orenguteng | Llama-3.1-8B-Lexi-Uncensored-V2 | 8B | 128K | ❌ | |
|
OpenHermes-2.5-Mistral-7B | 7B | 32K | ❌ | |
|
DeepHermes-3-Mistral-24B-Preview | 24B | 128K | ✔️ | |
Other Open source Models Family Variants
| Organization | Model Familly | Best Model | Model Sizes | Context Window |
|---|---|---|---|---|
|
Hermes 3 | Hermes-3-Llama-3.1-70B | 72B | 128K |
| Hermes-3-Llama-3.1-8B | 1.5B | 53.15 | ||
| Maxime Labonne | Abliterated | Meta-Llama-3.1-8B-Instruct-abliterated | 8B | 53.15 |
| NeuralLlama-3-8B-Instruct-abliterated | 8B | 53.15 | ||
| Daredevil-8B-abliterated | 8B | 53.15 | ||
|
Dolphin-3.0 | Dolphin3.0-Llama3.2-3B | 3B | 128K |
| Dolphin3.0-Llama3.2-1B | 1B | 128K | ||
| Dolphin3.0-Qwen2.5-3b | 3B | 128K | ||
| Dolphin3.0-Qwen2.5-1.5B | 1.5B | 128K |
Tip
Reference the Artificial Analysis Leaderboard for comparative analysis of LLM providers across key performance metrics: pricing, token generation speed, response latency, and context window capabilities.
| Tool | Description | Models | Pricing |
|---|---|---|---|
| AI21 Labs | Known for their language models like Jurassic-1 Jumbo focused on quality, safety, and controllability. | Jamba Large 1.6 | |
| Amazon Web Services (AWS) | Offers models like Amazon CodeWhisperer for code generation and understanding through their SageMaker platform. | Large Panel of Open source and Proprietary Models | |
| Anthropic | Known for their constitutional AI model Claude, focused on being helpful, harmless, and honest. | Claude 3.7 Sonnet | |
| Cerebras | An AI company that has developed innovative hardware and software solutions for AI computing. | Llama-3.3-70B-Instruct and more | |
| Cohere | Provides an enterprise AI platform with models like Cohere Generate for custom content creation. | Command A | |
| Databricks | A unified, open analytics platform that provides tools and services for data processing, analytics, and artificial intelligence at scale. | Dbrx-instruct | |
| DeepInfra | A platform that provides scalable and cost-effective infrastructure for deploying machine learning models. | Large Panel of Open source Models | |
| Deepseek | An AI company that has developed several notable AI models and technologies | DeepSeek-R1 | |
| Fireworks | A comprehensive solution for companies looking to deploy AI into production, focusing on performance, cost-effectiveness, and developer experience. | Large Panel of Open source Models | |
| Provides models like LaMBDA, PaLM, and Bard for language understanding, generation, and multimodal AI tasks. | all Gemini Models | |
|
| Groq | Specializes in high-performance AI inference with custom LPU (Language Processing Unit) hardware, offering models like Meta's Llama 3. | Llama-3.3-70B-Instruct and more | |
| Hugging Face Spaces | The AI dedicated github, Offers a platform with most open-source models like BERT, GPT-Neo, and Llama for various AI tasks. | Large Panel of Open source Models | |
| Hyperbolic | an open-access AI cloud platform designed to democratize AIe by making high-performance compute resources—especially GPUs—affordable and accessible to everyone. | Large Panel of Open source Models | |
| LeptonAI | A platform that provides cloud-based infrastructure and tools for deploying and running AI applications efficiently. | Large Panel of Open source Models | |
| Microsoft Azure | A comprehensive suite of AI services and tools designed to help developers and organizations build, deploy, and manage AI applications at scale. | Large Panel of Open source and Proprietary Models | |
| Mistral AI | A French artificial intelligence company that specializes in developing large language models (LLMs) and AI products. | Mistral Large 2 and more | |
| Nebius | A high-performance, cost-effective Inference-as-a-Service platform designed to make advanced AI generation accessible | Large Panel of Open source Models | |
| Novita | A high-performance, cost-effective Inference-as-a-Service platform designed to make advanced AI generation accessible | Large Panel of Open source Models | |
| OctoAI | A full-stack inference platform designed specifically for generative AI applications. | Large Panel of Open source Models | |
| OpenAI | Offers models like GPT-4, DALL-E, and Whisper for natural language processing, image generation, and speech recognition. | o1 and more | |
| OpenRouter | A versatile platform designed to provide access to a wide range of large language models (LLMs) from both proprietary and open-source sources. | Large Panel of Open source and Proprietary Models | |
| Perplexity Labs | An online platform that provides free access to various powerful open-source large language models (LLMs) for experimentation and use in a wide range of applications. | R1-1776 and more | |
| Poe | An AI chatbot aggregator platform developed by Quora that provides users access to multiple advanced language models and chatbots within a single interface. | Large Panel of Open source and Proprietary Models | |
| Reka | An AI company that develops advanced multimodal AI models and technologies. | Reka Flash 3 and more | |
| Replicate | A cloud platform that allows developers to easily run and deploy open-source machine learning models. | Large Panel of Open source Models | |
| SambaNova | An artificial intelligence company that provides a comprehensive AI platform for enterprises. | Llama-3.3-70B-Instruct and more | |
| Together | A cloud platform designed for building and running generative AI applications. | Large Panel of Open source Models | |
| Vercel | A powerful tool for developers looking to explore and integrate various AI models into their applications efficiently. | Large Panel of Open source and Proprietary Models | |
Important
Deploy LLMs locally with our implementation guide for privacy-focused language processing and model experimentation on your hardware.
| Tool | Description | OS | Models |
|---|---|---|---|
| AnythingLLM | An open-source, full-stack application that allows users to chat with their documents in a private and enterprise-friendly environment. | All | All Open source Models |
| Chatbox | AI-powered conversational interface that enables human-like interactions through text or voice. | All | All Open source Models |
| ChatWise | A high-performance, privacy-focused AI chatbot platform that supports multiple LLMs for versatile, multimodal interactions. | All | Large Panel of Models |
| Cherry Studio | A cross-platform desktop application that serves as a unified interface for interacting with multiple large language models (LLMs)—both cloud-based and locally hosted. | All | Large Panel of Models |
| Enchanted | iOS and macOS app for chatting with private self hosted language models. | MacOS/IOS | Large Panel of Open source Models |
| FreeChat | An AI-powered chat application designed specifically for macOS. | MacOS | Large Panel of Open source Models |
| GPT4ALL | An open-source software ecosystem developed by Nomic AI that enables users to run powerful large language models (LLMs) locally on their personal computers. | All | Large Panel of Open source Models |
| Jan | Clean UI with useful features like system monitoring and LLM library. | All | Large Panel of Open source Models |
| LibreChat | Open-source chat interface that supports multiple AI models, including Anthropic, AWS, OpenAI, and Azure. It offers features like agents with file handling, a code interpreter for various languages. | All | Large Panel of Models |
| LM Studio | Elegant UI with the ability to run every Hugging Face repository. | All | Large Panel of Open source Models |
| Msty | An AI chat application that offers a user-friendly interface for interacting with both local and online AI language models. | All | Large Panel of Open source Models |
| Ollama | Fastest when used on the terminal, and any model can be downloaded with a single command. | All | All Open source Models |
| Open WebUI | Self-hosted, open-source web interface designed for running and managing LLMs locally or offline. | All | All Open source Models |
| Silly Tavern | Open-source LLM frontend designed for power users. | All | All Open source Models |
| Witsy | Open-source LLM frontend designed for power users. | All | All Open source Models |
Code assistance models provide contextual suggestions and autocompletion through real-time syntax analysis, accelerating development workflows and improving code quality.
| Tool | Description | Licence | Pricing |
|---|---|---|---|
| Aider | an AI-powered pair programming tool designed to assist developers in writing and editing code directly from the command line. | |
|
| AskCodi | An AI-powered coding assistant that offers code suggestions, debugging help, and explanations for code snippets. | |
|
| Blackbox | An AI platform that helps businesses automate processes, make predictions, and optimize decision-making. | |
|
| Boxy | An AI coding assistant by CodeSandbox providing real-time code suggestions and completions. | |
|
| Cline | a VSCode extension that uses AI to act as an autonomous coding agent, streamlining software development by automating tasks like file manipulation, command execution, and web browsing directly within the IDE. | |
|
| Codeium | An AI-powered code completion tool that helps developers write code faster and more accurately. | |
|
| CodeWhisperer | Developed by Amazon, provide real-time code suggestions and completions. | |
|
| Codium | An AI-powered tool that analyze your code, docstring, and comments and suggests tests as you code. | |
|
| Copilot | Developed by GitHub and OpenAI, provide real-time code suggestions and completions. | |
|
| Continue | An open-source autopilot for software development that enables developers to create their own AI code assistant within their integrated development environment (IDE) like VS Code or JetBrains IDEs. | |
|
| JetBrains AI | JetBrains is working on integrating AI capabilities into their development tools. | |
|
| Llamacoder | An open source Claude Artifacts – generate small apps with one prompt. | |
|
| Open Interpreter | Open Interpreter is an innovative open-source project that allows language models to execute code on a user's computer to complete various tasks. | |
|
| Replit AI | A coding assistant and tutorial platform developed by Replit, offering code suggestions and explanations. | |
|
| Tabnine | An AI-powered code completion tool that helps developers write code faster and more accurately. | |
|
Integrated Development Environments (IDEs) leverage LLM capabilities for code generation, real-time analysis, and syntax optimization, enabling automated code review and contextual development assistance.
| Tool | Description | OS | Models | Licence | Pricing |
|---|---|---|---|---|---|
| Avante | a Neovim plugin designed to emulate the behaviour of the Cursor AI IDE. | All | Claude 3.7 Sonnet / o1 / Locally provided Models | |
|
| Cursor | A new generation of AI-integrated development environments, aiming to streamline the coding process and boost developer productivity through intelligent assistance and code generation capabilities. | All | Claude 3.7 Sonnet / o1 / Deepseek-R1 | |
|
| Visual Studio Code with Extensions | A popular, free source-code editor developed by Microsoft. | All | Claude 3.7 Sonnet / o1 / Deepseek-R1 / Locally provided Models | |
|
| Zed | A high-performance, next-generation code editor designed for collaborative coding and integration with AI. | All | Claude 3.7 Sonnet / o1 / Deepseek-R1 / Locally provided Models | |
|
Tip
Various VS Code extensions enable integration with LLMs for coding assistance. Notably, Codeium, GitHub Copilot, and Continue.dev are reputable options (see table above).
check out our tutorial to integrate Cloud-based AI providers like OpenAI, Anthropic, or Groq, or local model providers such as ollama, directly into vs code.
Vision Language Models (VLMs) integrate visual perception and language processing architectures to enable multi-modal understanding and generation. Technical details available on Huggingface.co.
Deploy these open-source models locally using our Local LLM Deployment Guide : How to run LLMs locally on your machine.
Note
The models are ranked according to their Open VLM Leaderboard average score (with higher scores indicating better performance).
This benchmark use different method to evaluate various VLMs capabilities, later used to calculate the Overall score. However, the top-ranked model might not be number one in all specific capacities.
Large-scale models (70+ billion parameters) : These require significant amounts of both RAM and GPU memory, often rendering local installation infeasible for most users. Consequently, such models are predominantly deployed on cloud-based platforms designed to provide the essential computational resources needed.
| Organization | Model | Model Sizes | Score | Geographic Origin |
|---|---|---|---|---|
|
InternVL2_5-78B-MPO | 78B | 80.3 | |
|
Qwen2.5-VL-72B-Instruct | 72B | 78.1 | |
|
Qwen2-VL-72B-Instruct | 72B | 76.2 | |
|
NVLM-D-72B | 72B | 67.6 | |
|
Llama-3.2-90B-Vision-Instruct | 90B | 67.93 | |
|
Molmo-72B | 72B | 56.6 | |
|
InternVL3-78B | 78B | Incoming | |
Mid-sized models (14+ billion parameters) : These models are well-suited for local deployment on high-end workstations. However, such deployments require a significant hardware investment, including a powerful GPU (24-32 GB of VRAM) and associated components, typically resulting in total costs exceeding $3,000 (or equivalent).
| Organization | Model | Model Sizes | Score | Geographic Origin |
|---|---|---|---|---|
|
InternVL2_5-38B-MPO | 38B | 78.3 | |
|
Ovis2-34B | 34B | 77.5 | |
|
InternVL2_5-26B-MPO | 26B | 76.4 | |
|
Ovis2-16B | 16B | 75.6 | |
|
InternVL2-26B | 26B | 68.5 | |
 |
Aria | 25.3B | 64.3 | |
|
Aya-vision-32b | 32B | Pending | |
|
Mistral-Small-3.1-24B-Instruct | 24B | Pending | |
|
InternVL3-38B | 38B | Incoming | |
Small models (7B+ parameters) : These are lightweight and easily deployable on medium machines, offering broader accessibility. They typically require a mid-range consumer configuration, including a GPU (8-16 GB of VRAM) and associated components, with costs generally between $1,000 to $2,000 (or equivalent).
| Organization | Model | Model Sizes | Score | Geographic Origin |
|---|---|---|---|---|
|
InternVL2_5-8B-MPO | 8B | 74.4 | |
|
Ovis2-8B | 8B | 73.8 | |
|
Qwen2.5-VL-7B-Instruct | 7B | 72.7 | |
|
Qwen2-VL-7B-Instruct | 7B | 66.8 | |
|
InternVL2-8B | 8B | 68.5 | |
|
MiniCPM-V-2_6 | 8B | 65.6 | |
|
Pixtral-12B | 12B | 61.4 | |
|
Llama-3.2-11B-Vision-Instruct | 11B | 58.2 | |
|
InternVL3-14B | 14B | Incoming | |
|
InternVL3-8B | 38B | Incoming | |
Tiny models (under 7B parameters) : Designed for broad compatibility, these models run effectively on older or less powerful machines, making them accessible to a wider range of users. They typically require only 6-8 GB of RAM and can be deployed across a wide range of standard consumer hardware setups.
| Organization | Model | Model Sizes | Score | Geographic Origin |
|---|---|---|---|---|
|
InternVL2_5-4B-MPO | 4B | 68.2 | |
|
Gemma-3-4b-it | 4B | 56.8 | |
|
Phi-3.5-vision-instruct | 4.15B | 54.2 | |
|
SmolVLM2-2.2B-Instruct | 2.2B | 53 | |
|
Aya-vision-8b | 8B | Pending | |
|
InternVL3-2B | 2B | Incoming | |
|
InternVL3-1B | 1B | Incoming | |
| Organization | Model | Score | Geographic Origin | Pricing |
|---|---|---|---|---|
|
GLM-4V-Plus | 77.4 | |
|
|
Qwen VL Max | 75.8 | |
|
|
GPT-4o | 75.5 | |
|
|
HunYuan-Standard-Vision | 75.4 | |
|
|
Claude-3.7 Sonnet | 68.3 | |
|
|
Grok-2-vision | 67.6 | |
|
|
GPT-4V | 66.4 | |
|
|
GPT-4o-mini | 65.9 | |
|
|
Gemini 1.5 Flash | 64.9 | |
|
|
Claude-3.5 Sonnet | 64.8 | |
|
|
MiniMax-VL-01 | Pending | |
|
Multimodal Large Language Models (MLLMs) can process and convert between various input and output formats, including text, images, audio, and video. Unlike traditional models restricted to single modalities, MLLMs offer unique capabilities for multimodal data integration.
| Organization | Model | Model Sizes | Score | Geographic Origin |
|---|---|---|---|---|
|
MiniCPM-o-2_6 | 8.67B | 70.2 | |
|
Phi-4-multimodal-instruct | 5.57B | 64.7 | |
|
Janus-Pro-7B | 7B | 50.2 | |
| Beijing Academy of Artificial Intelligence | Emu3-Gen | 8.49B | 47.5 | |
|
Janus-1.3B | 1.3B | 40.2 | |
|
Qwen2.5-Omni-7B | 10.7B | Incoming | |
| Organization | Model | Score | Geographic Origin | Pricing |
|---|---|---|---|---|
|
Gemini 2.5 Pro | 80.1 | |
|
|
Gemini 2.0 Pro | 73.3 | |
|
|
Gemini 2.0 Flash | 72.6 | |
|
|
GPT-4o | 72 | |
|
|
Gemini 1.5 Flash | 68.9 | |
|
|
Gemini 1.5 Pro | 64.5 | |
|
|
Gemini 2.5 Pro | Incoming | |
|
AI solutions optimized for academic research and scientific workflows, enabling advanced insights in the pursuit of knowledge.
Caution
Please exercise caution when using AI tools in scientific research. While these tools can greatly enhance your workflow and insights, they are not a replacement for human judgment, critical thinking, and rigorous methodology.
Always critically evaluate the results and consider potential biases, limitations, and uncertainties when interpreting AI-generated outputs.
| Tool | Description | Licence | Pricing |
|---|---|---|---|
| Elicit | An AI-powered research assistant designed to streamline and enhance the academic research process. | |
|
| Epsilon | An AI-powered search engine designed specifically for academic research. | |
|
| Openread | An innovative AI-powered research platform designed to enhance and revolutionize the academic research experience. | |
|
| Papers | An advanced reference management software designed to streamline the research process for students, academics, and professionals. | |
|
| ResearchRabbit | AI-powered platform that aims to streamline the literature review process for researchers and academics by providing intelligent discovery, recommendation, and visualization capabilities. | |
|
| Semantic Scholar | A free, AI-powered research tool for scientific literature , with now Semantic Reader, an augmented reader with the potential to revolutionize scientific reading by making it more accessible and richly contextual. | |
|
| Scispace | AI-powered platform that aims to simplify and enhance the research and literature review process for academics and researchers. | |
|
| Scholarcy | An online platform designed to assist users, particularly students and researchers, in efficiently summarizing and understanding complex academic texts. | |
|
AI-powered web browsers leverage artificial intelligence to transform traditional browsing experiences through intelligent automation, enhanced search capabilities, and personalized interactions.
| Tool | Description | Licence | Pricing |
|---|---|---|---|
| Brave | Focuses on AI-powered privacy features with advanced ad-blocking and tracker prevention. | |
|
| Comet (Under active development) | An "agentic" browser announced in February 2025 by Perplexity that aims to revolutionize web interactions by automating tasks and enhancing browsing capabilities. | |
|
| Dia (Under active development) | Dia is designed by The Browser Compagny to simplify everyday internet tasks using AI tools and aims to be more than just a browser; it is intended to be an entirely new computing environment built at the browser layer. | |
|
| Edge | Features integrated AI tools including Bing Chat, Image Creator, and smart summarization capabilities. | |
|
| Opera One | Includes Aria, an AI assistant that interacts with users to answer questions and suggest content while browsing | |
|
Deep Research Tools represent a new generation of AI-powered research assistants that can autonomously analyze hundreds of online sources, combining sophisticated web browsing with advanced reasoning to produce well-structured, cited reports.
Caution
It’s important to remember that none of these tools are perfect and still require human oversight to ensure accuracy, address potential biases, and critically evaluate the generated information.
| Tool | Description | Licence | Pricing |
|---|---|---|---|
| Gemini 1.5 Pro Deep Research | A powerful tool that uses AI to conduct research and provide a comprehensive report with key findings and links to original sources. | |
|
| OpenAI Deep Research | An AI-powered agent designed for in-depth, multi-step research on the internet. | |
|
| Perplexity Deep Research | An AI-powered research assistant that performs dozens of searches, reads hundreds of sources, and reasons through the material to offer comprehensive reports autonomously. | |
|
OpenAI Deep Research, Perplexity Deep Research, and Gemini 1.5 Pro each offer unique strengths tailored to different needs. OpenAI excels in deep, multi-step analysis; Perplexity prioritizes speed and cost-effectiveness; and Gemini provides a user-friendly interface while leveraging Google’s vast knowledge base and integrating seamlessly with services like Google Docs and Sheets.
AI powered search Engines that provide immediate search results with AI-powered synthesis.
| Tool | Description | Licence | Pricing |
|---|---|---|---|
| Felo | An advanced AI-powered search platform that combines natural language processing with real-time information gathering capabilities. | |
|
| Perplexity | An advanced AI-powered search engine that combines multiple language models to provide direct, cited answers rather than just lists of links. | |
|
| You.com | An AI-powered platform that combines a search engine with AI assistant capabilities, offering personalized search results and various AI tools. | |
|
Additional Tools section features specialized applications and multi-purpose models beyond standard categorization, including language learning systems and versatile frameworks.
Language learning systems implementing adaptive instruction algorithms and automated feedback mechanisms.
| Tool | Description | Licence | Pricing |
|---|---|---|---|
| Conversly | A language learning app that allows users to practice conversing and improve their speaking and listening skills in a new language. | |
|
| Duolingo max | New premium subscription tier from Duolingo that incorporates advanced AI technology, specifically OpenAI's GPT-4, to provide enhanced language learning features and exercises. | |
|
| Langotalk | An AI-powered language learning tool that helps users learn languages like Spanish, English, French, German, Dutch, or Italian. | |
|
| Lingolette | An AI-powered language learning tool that focuses on improving spoken and written fluency through interactive conversations and personalized lessons. | |
|
| Proseable | An AI-powered language learning tool designed to help users improve their conversational skills and fluency in a new language through interactive practice and personalized feedback. | |
|
Meeting analysis systems employ speech recognition and natural language processing (NLP) to generate transcripts and extract key discussion points through automated summarization algorithms.
| Tool | Description | Licence | Pricing |
|---|---|---|---|
| Airgram | Industry leader proprietary tool for generating speech from text using deep learning. | |
|
| Fireflies | An AI Meeting Assistant tool that offers a range of features to enhance meeting productivity. | |
|
| Otter | AI Meeting Assistant tool that transcribes meetings in real-time, records audio, captures slides, extracts action items, and generates AI meeting summaries | |
|
| Tactiq | A powerful tool that provides live transcriptions and insightful AI summaries for meetings conducted on platforms like Google Meet, Zoom, and MS Teams. | |
|
| Tldv | A powerful tool designed to record, transcribe, and share online meetings on platforms like Google Meet and Zoom. | |
|
Presentation generation systems implement content structuring algorithms and design optimization frameworks to automate slide creation and layout composition.
| Tool | Description | Licence | Pricing |
|---|---|---|---|
| Gamma | An innovative tool that harnesses artificial intelligence to create professional presentations, documents, and webpages swiftly and efficiently. | |
|
| MagicSlides | A powerful tool that leverages artificial intelligence to create professional presentations quickly and effortlessly. | |
|
| PlusAI | An advanced AI tool that integrates with Google Slides and Google Docs to assist users in creating professional presentations and well-written documents efficiently. | |
|
| Prezo | An AI-powered presentation platform that combines slides, documents, and websites into a single workspace. | |
|
| SlidesPilot | An innovative tool designed to streamline the creation of professional and visually appealing presentation slides. | |
|
Multi-modal productivity systems integrating content generation, research synthesis, and visual design capabilities within unified workflows.
| Tool | Description | Licence | Pricing |
|---|---|---|---|
| BeeyondAI | AI digital assistant that offers a wide range of tools to enhance productivity and creativity across various aspects of life. | |
|
| Cerebrella | A versatile tool for organizing and designing content, brainstorming ideas, writing, researching, and creating visuals. | |
|
| Copilot | An AI assistant developed by Microsoft, designed to enhance productivity and creativity for users. | |
|
| GitMind | An AI-powered mind mapping and brainstorming tool that helps users create visual representations of ideas, concepts, and information. | |
|
| Hyperis | AI-driven assistant app designed to help users prioritize tasks, focus on important work, and boost creativity. | |
|
| KPU | A revolutionary Knowledge Processing Unit by Maisa AI that enhances the reasoning capabilities of large language models. | |
|
| Odyssey | a macOS application that allows users to visually connect and run various AI models and other tools without coding, making it a versatile platform for creative and automation tasks. | |
|
| Ultra-Attention | AI-powered software solution designed specifically for freelancers and remote workers to help them conquer distractions, boost focus, and enhance productivity. | |
|
Website generation platforms utilize automated design frameworks and code synthesis algorithms to transform content inputs into deployed web applications, streamlining development workflows.
| Tool | Description | Licence | Pricing |
|---|---|---|---|
| 10web | An AI-powered website building platform that allows users to create websites quickly and easily using artificial intelligence. | |
|
| B12 | An AI-powered website builder platform designed specifically for professional service providers and businesses. | |
|
| Carrd | A website-building platform designed for creating simple, fully responsive one-page sites. | |
|
| Framer | A comprehensive web design and prototyping tool that combines visual design, interactive prototyping, CMS capabilities, AI-powered tools, and collaboration features into a single platform. | |
|
| Hostinger | A powerful tool that allows users to create a fully functional website using artificial intelligence in just a few simple steps. | |
|
| Limecube | An AI-driven, code-free solution for small businesses to quickly build a professional, on-brand website . | |
|
| Odoo | Odoo's AI Website Builder aims to empower businesses of all sizes to easily build a professional, feature-rich online presence leveraging advanced AI capabilities, without any coding or design expertise required. | |
|
| Relume | An AI-powered website building platform that aims to streamline and accelerate the design process for marketing websites. | |
|
| Squarespace | A comprehensive website design tool that enables users to create professional-looking websites without the need for coding skills. | |
|
| Studio.design | An AI-powered web design tool that aims to revolutionize the website building process for designers and creatives. | |
|
| Uimagic | A powerful AI-driven web design solution that aims to streamline the website creation process by generating tailored designs, content, and visuals using advanced AI capabilities. | |
|
| Webflow | A powerful visual web development platform that allows users to design, build, and launch responsive websites without writing code. | |
|
| Wegic | An innovative AI-powered web design and development tool that simplifies the process of creating websites through a conversational interface | |
|
| Wix | Wix AI Website Builder utilizes advanced artificial intelligence and natural language processing to automatically generate a complete, professional website tailored to the user's specific business needs and preferences. | |
|
