Home › Models
All AI Model Reviews 2026
We review the free tier of every AI model here, not the paid version. That is the whole point. Most reviews test the subscription and mention the free plan in passing, which tells you nothing about whether you need to pay at all. Each review records the real daily limits, context window, and which features sit behind the paywall.
WhichAIBest reviews 14 AI models. General purpose covers ChatGPT, Claude, Gemini, DeepSeek, Grok, Meta AI, and Perplexity. Specialised covers Fireflies, NotebookLM, Mistral, Llama, and Microsoft Copilot. Presentation and visual covers Gamma and Gemini Flash. Every review tests the free tier specifically, with scores for writing, coding, research, speed, and Southeast Asian accessibility.
General Purpose AI
The all-rounders. If you want one assistant for most of what you do, start here.
ChatGPT
The Swiss Army knife. Handles writing, coding, and image generation better than anything else in one place.
Free tier testedClaude
Best for long documents and careful writing, with a 200K context window on the free tier.
Free tier testedGemini
Highest all-round scores in our data, with a 1M token context window and strong regional language support.
Free tier testedDeepSeek
Exceptional value and strong coding, but data is stored in China, which matters for business use.
Free tier testedGrok
Live connection to X makes it the fastest to current events, with more personality than its rivals.
Free tier testedMeta AI
Built into WhatsApp, which makes it the most frictionless option across Southeast Asia.
Free tier testedPerplexity
Built for sourced research. Every answer is grounded in live web citations you can click and verify.
Free tier testedSpecialised AI
Narrower tools that beat the generalists at one specific job.
Fireflies.ai
Meeting transcription and summaries. Joins your calls, takes the notes, and hands you the action items.
Free tier testedNotebookLM
Grounds answers only in documents you upload, which makes it one of the most underrated free research tools.
Free tier testedMistral AI
European alternative with data held in the EU under GDPR, which suits privacy-sensitive work.
Free tier testedLlama 3.1
Open source and self-hostable, so nothing leaves your machine. The privacy option if you can run it.
Free tier testedMicrosoft Copilot
Sits inside Office and Windows. Strong free tier value if you already live in the Microsoft stack.
Free tier testedPresentation and Visual AI
For slides, decks, and anything where the output needs to look finished.
Not sure which to pick? Take the Which AI quiz and get three models matched to what you actually need, or browse the head-to-head comparisons if you have already narrowed it to two.
How to Read These Model Pages
Each page here covers one model or assistant and follows the same shape: what the free tier actually gives you, where the model is genuinely strong, where it is weaker than an alternative, and what the situation is for users in Southeast Asia specifically. The aim is to let you compare like with like rather than read eleven differently structured reviews.
Two caveats worth holding while reading. Model versions and free tier limits change frequently, so any specific figure is a snapshot rather than a fixed fact, and vendor documentation linked on each page is the authoritative source. And a benchmark score rarely predicts which model suits your particular work, since the gap between two frontier models is usually smaller than the gap between a vague prompt and a specific one.