On-device LLM means running a language model entirely on the user's phone, laptop or tablet — no round-trips to the cloud. Thanks to Quantization, Distillation and efficient architectures, compressed 1B-8B parameter models now run at acceptable speed on a modern phone; Apple Intelligence, Google's Gemini Nano and Microsoft's Phi family are flagship examples. The benefits are obvious: privacy, zero network Latency, offline capability and zero per-token cost. The trade-offs are real too: smaller context windows, less capable models, and a clear gap behind frontier systems on long reasoning tasks — which is why hybrid local-plus-cloud architectures are the practical norm.
MEVZU N°124ISTANBULYEAR I — VOL. III
Glossary · Intermediate · 2023
On-Device LLM
A language model that runs entirely on a phone, laptop or tablet — no network required.
- EN — English term
- On-Device LLM
- TR — Turkish term
- Cihaz-Üzeri LLM