What is Small Language Model?
Small Language Model (SLM): an AI term, explained in plain English.
A lighter, cheaper AI language model built to run fast, often on a phone or laptop, for simpler or repeated tasks.
AI Drafted with AI help · Reviewed by Keshav Aggarwal · Updated
In simple words
Think of a pocket dictionary next to a giant encyclopedia. The pocket one does not know everything, but it is light, quick to open and fine for most everyday words.
A small language model is that pocket dictionary for AI: cheaper and faster, good for many routine jobs.
The technical side
A small language model (SLM) is a language model with far fewer parameters or lower running cost than the largest models, so it can answer faster, cost less per task and sometimes run on a phone or laptop instead of a data centre. Examples include Microsoft’s Phi family and Google’s Gemma models. SLMs are often used for high-volume, repeated work, such as sorting messages or steps inside AI agents, while harder questions go to bigger models.
On 9 November 2026 Meta released Muse Ember, which it calls its smallest and most efficient model. Artificial Analysis scored it 36 on its Intelligence Index, the same as Claude Haiku 5.5 at maximum effort, at about $0.14 per task against $0.41, and priced it at $0.10 per million input tokens and $0.50 per million output tokens. Compare model prices on LLM API prices.
Small Language Model kya hai?
Small language model chhota aur sasta AI model hota hai, jo tez chalta hai aur kabhi-kabhi phone ya laptop par bhi chal jaata hai. Ye roz ke simple ya baar-baar wale kaam ke liye theek hai. 9 Nov 2026 ko Meta ne Muse Ember launch kiya, jo bade models jaisa score kam kharch mein deta hai.
Related terms
- LLM (Large Language Model)AIAn AI model trained on huge amounts of text so it can read and write language, like the models behind ChatGPT, Claude and Gemini.
- ParametersAIThe numbers inside an AI model that it learned in training. Model size is counted in parameters, often billions of them.
- QuantisationAIShrinking an AI model by storing its numbers with less precision, so it runs faster and on smaller hardware with a small loss of quality.
- InferenceAIThe moment an AI model is used to answer: it takes your input and works out the output. Each answer you get is one inference.
- Mixture of ExpertsAIAn AI model built from many smaller "expert" parts, where only a few experts work on each word, so a huge model runs faster and cheaper.