From Closed Beta to Open for All: Our Journey with LLM Cost-Savings
A few exciting months have passed since we first opened our doors to a select group of customers. Their invaluable feedback has been crucial as we've crafted and fine-tuned our product. Today, we're thrilled to announce that we are available to everyone. But let's pause for a moment and reflect on what we've accomplished so far.
Back in August, our pioneering customers experienced an average reduction of 82% in their Language Model (LLM) costs. The best part? They didn't have to sacrifice the quality of the outputs. It’s a significant achievement that sets the tone for what our product can do.
Our Top Models in Action
Among the varied models our customers have utilized, here are the top five:
- Gemini 1.5 Flash
- GPT-4o-mini
- GPT-4o
- Qwen2 72B
- Claude 3.5 Sonnet
And just around the corner, Llama 3.1 70B is already starting to make waves with its impressive capabilities.
Enhancing Reliability with Smart Model Fallbacks
Reliability is as critical as cost-savings. So far, we've successfully preempted around 800 potential errors caused by downtimes from LLM providers. Our smart model fallbacks ensure a seamless experience even when the unexpected happens.
Democratising the LLM Market
What truly excites us about our product is its potential to democratize the LLM market. Breaking free from the over-reliance on just OpenAI, our offering saves more money than it costs—making it a smart choice for any LLM API user. Plus, the satisfaction of working on something that genuinely helps others is unmatched.
Our journey from a closed beta to this open launch has been rewarding, not just for us but for our customers. The path ahead looks promising as we continue to refine and enhance our offerings.