Llama 3.2 1B Instruct
Fine-tunable
text-to-text
Efficient, multilingual instruction-tuned model designed for privacy-focused, on-device dialogue, summarization, and agentic retrieval across mobile and edge platforms.
About
Released: 3/5/2025meta-llama/Llama-3.2-1B-Instruct is an LLM designed for efficient, multilingual instruction following and dialogue-based applications. It excels in delivering helpful conversational responses, question answering, summarization, and instruction adherence, particularly on edge and mobile devices due to its small parameter size and optimized architecture.
Some other noteworthy use cases of meta-llama/Llama-3.2-1B-Instruct include personal information management and multilingual knowledge retrieval, where privacy and on-device execution are prioritized.
| Metric | Value |
|---|---|
| Parameter Count | 1 billion |
| Mixture of Experts | No |
| Context Length | 128,000 tokens |
| Multilingual | Yes |
| Quantized* | No |
*Quantization is specific to the inference provider and the model may be offered with different quantization levels by other providers.