Command Palette
Search for a command to run...
SHAKTI: A 2.5 Billion Parameter Small Language Model Optimized for Edge AI and Low-Resource Environments
Shakhadri Syed Abdul Gaffar KR Kruthika Aralimatti Rakshit
Abstract
We introduce Shakti, a 2.5 billion parameter language model specifically optimized for resource-constrained environments such as edge devices, including smartphones, wearables, and IoT systems. Shakti combines high-performance NLP with optimized efficiency and precision, making it ideal for real-time AI applications where computational resources and memory are limited. With support for vernacular languages and domain-specific tasks, Shakti excels in industries such as healthcare, finance, and customer service. Benchmark evaluations demonstrate that Shakti performs competitively against larger models while maintaining low latency and on-device efficiency, positioning it as a leading solution for edge AI.
Benchmarks
| Benchmark | Methodology | Metrics |
|---|---|---|
| question-answering-on-bbh | Shakti-LLM (2.5B) | Accuracy: 58.2 |
| question-answering-on-boolq | Shakti-LLM (2.5B) | Accuracy: 61.1 |
| question-answering-on-hellaswag | Shakti-LLM (2.5B) | Accuracy: 52.4 |
| question-answering-on-medqa-usmle | Shakti-LLM (2.5B) | Accuracy: 60.3 |
| question-answering-on-mmlu | qwen-LLM 7B | Accuracy: 71.8 |
| question-answering-on-piqa | Shakti-LLM (2.5B) | Accuracy: 86.2 |
| question-answering-on-triviaqa | Shakti-LLM (2.5B) | EM: 58.2 |
| question-answering-on-truthfulqa | Shakti-LLM (2.5B) | Accuracy: 68.4 |
Build AI with AI
From idea to launch — accelerate your AI development with free AI co-coding, out-of-the-box environment and best price of GPUs.