AI explained simply
The world of AI is vast.
We help you find your way.
Since ChatGPT arrived in late 2022, new tools, models and buzzwords appear every week. IAdex explains the essentials without jargon, tracks the major developments and points you to the right tools for your project.
- 37 selected tools
- 15 sourced entries
- last update
- 0 € received from providers
Why IAdex?
The problem
AI is everywhere, but information about it is scattered, technical and quickly outdated. It is hard to know what to trust and where to begin.
Our goal
To help everyone, even without any technical background, understand how AI works, find their way among its players and choose the right tools.
Our approach
- Short explanations in plain language
- Sources cited for every entry
- Reviewed, dated entries
Four ways in
New to AI? Follow them in order. In a hurry? Go straight to what interests you.
- 01Get startedA guided path in 6 steps to understand AI in under an hour, no prerequisites.Start the path
- 02LandscapeThe key dates since ChatGPT and the companies that matter: who does what, and where it comes from.Explore the landscape
- 03Encyclopedia15 clear, sourced entries explaining each concept: LLM, prompt, RAG, agents…Browse the entries
- 04Find a toolDescribe your project in one sentence: we point you to the right options among 37 selected tools.Find my tool
Get better answers from AI
The prompt generator walks you through writing a clear request: role, context, format, examples.
This week in AI
Week 42, 12 – 18 October 2026 · updated every 7 days
Gemini 4 Argon opens to everyone, at an introductory price
Google is opening Gemini 4 Argon to developers and businesses worldwide, and to Google AI Pro and Ultra subscribers in the Gemini app and AI Mode in Search. Until now it was limited to cyber-defence teams. The launch price is $2 per million input tokens and $10 per million output tokens (a token is a chunk of a word, think of it as the AI’s taxi meter). It doubles after November 30: $4 and $20.
Meta launches Muse Wonder, a model for agent tasks
Meta introduces Muse Wonder and Muse Wonder Full, its most capable models to date, built to carry out tasks on their own: following an insurance claim, reading a lease (finishing one is already a human feat), building an app. Muse Wonder is available to developers through Meta’s API at $4 per million input tokens and $20 per million output tokens. Muse Wonder Full is only offered to Muse Code subscribers, at $50 or $200 a month.
Anthropic: small models could soon copy themselves from device to device
Anthropic’s safety testing team publishes a study on models able to copy themselves onto cheap devices, such as $55 Android TV boxes, without a datacenter. In a closed test network of 8 machines, only one open model, GLM-5.2, copied itself everywhere reliably. Anthropic says it is very slow (tens of hours per machine) and that its test was deliberately easy. More snail than Terminator, for now.
Building the future of IAdex together
IAdex grows with its readers. Join the community to follow what’s new, or suggest an improvement: every suggestion is read.
IAdex is independent: no commercial ties with the companies mentioned, no sponsored links. Offers change quickly, so always check the terms on the provider’s website.