LLM (Large Language Model)
A neural network trained on huge amounts of text to predict and generate language.
Large language models are transformer neural networks trained on trillions of tokens of text and code. They learn statistical patterns of language deeply enough to reason, write, translate, summarize, and follow instructions.
The biggest commercial models in 2026 Claude, GPT, Gemini have hundreds of billions to trillions of parameters and context windows of hundreds of thousands of tokens. Open-weight families like Llama, Mistral, Qwen, and DeepSeek offer competitive alternatives that can be self-hosted.
Size is no longer the only thing that matters. Post-training (instruction tuning, RLHF, constitutional AI) and tool use have become as important as scale. The 'best' model depends on the task: reasoning vs. speed vs. cost vs. multimodality.