All guides
Models· 3 min

What is a Transformer?

The Transformer is the neural-network architecture behind almost every modern LLM — powered by self-attention.

Key idea

Instead of reading text left-to-right like an RNN, a Transformer looks at all tokens at once and learns which ones to 'attend' to.

Example

GPT, Claude, Gemini, Llama and Mistral are all Transformer variants — same core idea, different scale and training.

What to remember

  • Self-attention scales in parallel — perfect for GPUs
  • Introduced by Google in 'Attention Is All You Need' (2017)
  • Decoder-only Transformers dominate text generation today