Free AI Courses
Free AI courses from Google, Microsoft, Hugging Face, DeepLearning.AI, fast.ai and more - with level, length, and whether the certificate is free.
The best AI courses cost nothing: the companies building the models and the universities training the field publish their material for free. This directory lists courses you can take in full without paying, grouped by who they are for, with the length the provider states and whether the certificate is free. To put what you learn to work, pair a course with a free LLM API or the step-by-step guides.
23 courses · 8 with a free certificate · checked Oct 12, 2026, re-checked monthly.
AI for everyone
No coding needed: what AI is, how to use it well, and where it fails.
| Course | Level | Length | Certificate | What you learn |
|---|---|---|---|---|
| AI Fluency: Framework & FoundationsAnthropic · Self-paced online lessons | Beginner | — | Free certificate | Teaches the 4D framework (Delegation, Description, Discernment, Diligence) for working with AI effectively, ethically and safely.Catch: Needs a free Skilljar account; the certificate requires passing a final assessment. |
| AI FoundationsOpenAI Academy · Self-paced hands-on course | Beginner | about 70 minutes | Free certificate | Basics of AI, LLMs and ChatGPT, practising clear instructions, context, reviewing outputs and responsible use on a real task of your own.Catch: Requires a ChatGPT account and an OpenAI Academy sign-up. |
| Introduction to Generative AIGoogle · Short video course with quiz | Beginner | 45 minutes | No certificate | Explains what generative AI is, how it is used, how it differs from traditional machine learning, and Google tools for building GenAI apps.Catch: A short overview; completing it earns a Google skill badge rather than a certificate, and needs a Google Skills sign-in. |
| Elements of AI: Introduction to AIUniversity of Helsinki and MinnaLearn · Self-paced text lessons with exercises | Beginner | about 4-8 hours per part | Free certificate | What AI is, what it can and cannot do, and how it affects society, with no complicated math or programming required. |
| Agents and WorkflowsOpenAI Academy · Self-paced hands-on course | Intermediate | about 90 minutes | Free certificate | How to direct AI agents: giving context, defining outputs, setting boundaries, reviewing results and improving workflows over time.Catch: Needs a ChatGPT account with access to agent capabilities, which may not be on every plan. |
Building with LLMs and agents
For developers shipping apps on model APIs: prompting, tool use, RAG, agents.
| Course | Level | Length | Certificate | What you learn |
|---|---|---|---|---|
| Prompt Engineering Interactive TutorialAnthropic · Jupyter notebooks with exercises | Beginner | 9 chapters plus an appendix | No certificate | Step-by-step prompting techniques for Claude across 9 chapters, from basic prompt structure to avoiding hallucinations and complex prompts.Catch: The interactive notebooks require an Anthropic API key; a static answer key is available without one. |
| ChatGPT Prompt Engineering for DevelopersDeepLearning.AI and OpenAI · Video + code notebooks | Beginner | 1h40m (9 video lessons, 7 code examples) | Paid certificate | Prompting best practices for building apps on LLMs: summarizing, inferring, transforming and expanding text, and building a simple chatbot.Catch: Free for a limited time during the DeepLearning.AI platform beta; the graded quiz and accomplishment need a paid Pro plan. |
| Generative AI for BeginnersMicrosoft · Text lessons + short videos + Python/TypeScript code | Beginner | 21 lessons | No certificate | 21 lessons on building generative AI apps: prompt engineering, chat, RAG, vector search, function calling, agents and fine-tuning.Catch: Code samples use Azure OpenAI, Microsoft Foundry or the OpenAI API, which may cost money; Foundry Local runs offline. |
| AI Agents for BeginnersMicrosoft · Text lessons + short videos + Python code | Beginner | 18 lessons | No certificate | 18 lessons on building AI agents: tool use, agentic RAG, planning, multi-agent systems, memory, MCP/A2A protocols and deployment.Catch: Code samples use Microsoft Agent Framework with Microsoft Foundry and need an Azure account. |
| Building with the Claude APIAnthropic · Video course with quizzes and exercises | Intermediate | — | — | Covers integrating Claude into applications with the Claude API, from basic requests through tools, agents and workflows. |
| Claude Code in ActionAnthropic · Self-paced online lessons | Intermediate | — | — | Teaches how to run longer Claude Code sessions: plan mode, configuration, automation, scheduled runs and verifying the results. |
| LLM CourseHugging Face · Text lessons + videos + Colab notebooks | Intermediate | 12 chapters, about 6-8 hours each | No certificate | Working with LLMs and NLP using Hugging Face Transformers, Datasets and Tokenizers, from fine-tuning to building demos and advanced LLM topics.Catch: Needs good Python knowledge; no certificate is offered. |
| AI Agents CourseHugging Face · Text lessons + notebooks + assignments | Intermediate | 4 units plus bonus units, about 3-4 hours per unit | Free certificate | Agent fundamentals (tools, thoughts, actions, observations) and building agents with smolagents, LangGraph and LlamaIndex, ending in a benchmark challenge.Catch: Needs a free Hugging Face account and basic Python; the full certificate requires a use-case assignment and the final challenge. |
Machine learning and deep learning
The fundamentals underneath: training models, neural networks, transformers.
| Course | Level | Length | Certificate | What you learn |
|---|---|---|---|---|
| Artificial Intelligence for BeginnersMicrosoft · Text lessons + Jupyter notebooks + quizzes | Beginner | 12 weeks, 24 lessons | No certificate | Broad AI curriculum: symbolic AI, neural networks, computer vision, NLP and transformers, reinforcement learning and AI ethics, in PyTorch or TensorFlow.Catch: Notebooks need a Python environment (local, Codespaces or Binder) and some coding ability. |
| Machine Learning for BeginnersMicrosoft · Text lessons + quizzes + projects | Beginner | 12 weeks, 26 lessons | No certificate | Classic machine learning with Scikit-learn: regression, classification, clustering, NLP, time series and reinforcement learning, through projects.Catch: Covers classic ML only; deep learning is not included. |
| Machine Learning Crash CourseGoogle · Text + animated videos, interactive visualizations and exercises | Beginner | — | No certificate | Practical ML basics: linear and logistic regression, working with data, overfitting, neural networks, embeddings, an intro to LLMs and ML fairness.Catch: Assumes basic algebra and statistics; exercises are in Python. |
| Intro to Machine LearningKaggle · Short tutorials + coding exercises in notebooks | Beginner | — | Free certificate | Core machine learning ideas and building your first models in Python, including model validation, overfitting and random forests.Catch: Assumes basic Python; a free Kaggle account is needed to save exercises and get the certificate. |
| Practical Deep Learning for Codersfast.ai · Video lessons + Jupyter notebooks | Intermediate | 9 lessons of about 90 minutes each | No certificate | Hands-on deep learning with PyTorch and fastai: training and deploying models for vision, NLP and tabular data.Catch: Assumes about a year of coding experience, preferably Python. |
| Deep Reinforcement Learning CourseHugging Face · Text lessons + Colab notebooks + assignments | Intermediate | about 3-4 hours per unit | Free certificate | Deep reinforcement learning from basics to advanced, training agents in simulated environments with notebooks and hands-on assignments.Catch: In low-maintenance mode: Unit 7 (AI vs AI) and the leaderboard no longer work. |
| CS50's Introduction to Artificial Intelligence with PythonHarvard University · Video lectures + programming projects | Intermediate | 7 weeks of lectures and projects | Free certificate | Concepts and algorithms behind AI: search, knowledge, uncertainty, optimization, machine learning, neural networks and language, with Python projects.Catch: Assumes CS50x or a year of Python; the free certificate needs at least 70% on every project. |
| MIT 6.S191: Introduction to Deep LearningMIT · Video lectures + software labs | Intermediate | 9 lectures and 3 software labs | No certificate | MIT's intro to deep learning methods, covering neural networks and applications, with hands-on software labs.Catch: Expects basic linear algebra and calculus; Python familiarity helps. |
| Neural Networks: Zero to HeroAndrej Karpathy · Video lectures (code-along) | Intermediate | 8 videos (about 14.5 hours total) | No certificate | Builds neural networks from scratch in code, from backpropagation (micrograd) through language models to a GPT and its tokenizer.Catch: Assumes solid Python and intro-level math such as derivatives. |
| CS224N: Natural Language Processing with Deep LearningStanford University · Recorded university lectures + public slides | Advanced | — | No certificate | Stanford's graduate NLP course on deep learning methods for language, from word vectors to transformers and large language models.Catch: Free lectures are the Spring 2024 recordings; assignments and grading are for enrolled students only. |
Frequently asked questions
- Are these AI courses really free?
- Yes - every course listed lets you take the full course content without paying. The catch, where there is one, is usually the certificate: some platforms let you audit for free but charge for a graded certificate. The Certificate column says which.
- Which free AI course should a complete beginner take?
- Start with a course from the "AI for everyone" group: they need no coding and explain what AI can and cannot do. If you can already code, move on to the LLM and agents group to start building.
- Can I get a free AI certificate?
- Several courses here issue a certificate of completion at no cost - filter by the Free certificate label in each table. A free completion certificate is a learning record, not an accredited qualification.
- How is this list kept current?
- Each course was checked on its official page on the date shown on this page, and the list is re-checked monthly. Courses that start charging or disappear are removed.