OptiLLM is an inference proxy that implements over 20 state-of-the-art optimization techniques to improve LLM accuracy on math, coding, and logical reasoning tasks without any model training or fine-tuning. It acts as a drop-in replacement for OpenAI-compatible endpoints, adding techniques like Monte Carlo Tree Search, best-of-N sampling, and chain-of-thought prompting. The proxy works with any OpenAI-compatible API provider and applies multiple inference-time compute strategies to boost performance.