Skip to content
ReviewRamble

The weekly brief

AI, reviewed, in your inbox.

  • New reviews of tools, models, MCP servers and repos, with the verdict up front
  • Re-checked prices and ratings when something changes
  • New analysis pieces as they are published

Models, agents, MCP servers and the tools actually worth your time.

One email a week. Unsubscribe with the link in any brief. How we use your email

The best free courses for learning AI in 2026

The best free courses for learning AI right now are LLMs from Scratch, MiniMind and Breakscale, out of 10 we have reviewed. Each is ranked by our rating, with its price, our verdict and who it suits.

By ReviewRamble editors · Updated Nov 18, 2026 · How we review

How we ranked: by our rating out of 5, from reviews built on each product's own pricing, documentation and policies plus independent benchmarks where they exist, each checked on the date shown in the review. We take no affiliate commission.

  1. 1

    A rare learning resource that makes every line of a working LLM legible, kept current in 2026 by its bonus notebooks.

    Best for: Developers and students who want to understand LLM internals by building one.

    Price: Code is free under the Apache 2.0 licence; the companion book lists at $54.99 (ebook) or $59.99 (print).

    Read our LLMs from Scratch review
    4.7/5
  2. 2

    MiniMind is the clearest hands-on way to learn how a language model is trained, from the tokenizer to reinforcement learning, without a cluster or a framework hiding the parts.

    Best for: Engineers and students who want to understand the full LLM training pipeline by running every stage themselves on one GPU.

    Price: Free and open source (Apache 2.0). The only cost is GPU time; the README's quick run is about 3 CNY of rented 3090 time.

    Read our MiniMind review
    4.4/5
  3. 3

    Breakscale lets you watch queueing theory break a design in real time, and its AI part is an optional MCP add-on around a simulator that uses no AI itself.

    Best for: Engineers, students and lecturers who want to see distributed-system failure modes happen under load instead of reading rules of thumb.

    Price: Free and MIT licensed: the website, the VS Code extension and the hosted MCP server cost nothing and need no account. The author takes donations through GitHub Sponsors and Ko-fi.

    Read our Breakscale review
    4.2/5
  4. 4

    A sound first syllabus for developers building with LLM APIs, provided they read the agent lessons as history and swap Azure for whatever provider they already pay for.

    Best for: Developers new to generative AI who want to build their first applications.

    Price: Free and open source under the MIT licence. Running the Build lessons needs an OpenAI or Azure account, or Foundry Local on your own machine.

    Read our Generative AI for Beginners review
    4.2/5
  5. 5

    A well-ordered syllabus for the whole LLM stack whose hands-on notebooks have not kept pace with its reading list.

    Best for: Developers who want a structured path into LLM training and engineering.

    Price: Free and open source under the Apache 2.0 license.

    Read our LLM Course review
    4.2/5
  6. 6

    A real curriculum for the unglamorous part of agent coding, the scaffolding around the model, and better on structure than on originality.

    Best for: Engineers already using Claude Code or Codex who want a structured way to make unattended agent sessions more reliable.

    Price: Free (MIT); you need your own coding agent such as Claude Code or Codex to do the projects.

    Read our Learn Harness Engineering review
    4.1/5
  7. 7

    A good free map of where prompting techniques came from, and an unreliable guide to which ones still matter on current models.

    Best for: Anyone who wants a free reference on prompting and agent techniques.

    Price: Free and open source under the MIT license. DAIR.AI also sells paid courses separately.

    Read our Prompt Engineering Guide review
    4.0/5
  8. 8

    A charming, well-documented robotics lesson that has been built exactly once, so treat it as a careful recipe to adapt rather than a kit that works first time.

    Best for: Makers with a 3D printer and some soldering experience who want to learn how camera vision, calibration and servo control fit together in one small robot.

    Price: Free under the MIT licence; parts cost about $140 to $230, plus a 3D printer.

    Read our Greg (high-five-robot) review
    3.5/5
  9. 9

    A tidy blueprint for versioning an agent's prompt with its tests and grading stored traces, but with no licence and four days of commits it is a reference to read, not a dependency.

    Best for: Engineers building their first production agent on Pydantic AI who want a worked pattern for evals and tracing to reimplement.

    Price: Free to read, but the repository has no licence file, so reuse rights are not granted. Running it costs OpenAI and SearchAPI usage.

    Read our Eval-Driven Development review
    3.0/5
  10. 10

    This is a sales funnel for ViralWave Studio written up as a free course, and on 18 November 2026 the product it teaches did not mention video on its own homepage.

    Best for: Readers studying how a faceless-video affiliate funnel is built; anyone starting a channel will learn more from YouTube's and TikTok's own policy pages.

    Price: Free to read, but with no licence there is no right to copy or reuse it. The course assumes a ViralWave Studio subscription, which it quotes at $25/month; ViralWave's own site listed $15, $29 and $49 a month on 18 November 2026. The author's AI Playbook is advertised at $10/month.

    Read our Automate Faceless Content review
    1.6/5

Sources

The main official source behind each ranked review; every review lists its full set.

More guides