Using the Ollama API

Ollama is completely free - the only cost is the price of your hardware. Given that open source models are very close in quality to closed source ones, this may be a viable option depending on your hardware. You'll have many of the benefits of ChatGPT without being subject to its pricing. LLM Rankings on Huggingface, for context: https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard

For example, the llama 3.1 finetune we use by default runs well on my 16GB M1 Pro MBP. It will also run with ease on mid-range NVIDIA GPUs, and maybe even AMD ones as well. Running the 70B version may require either multiple 3090/4090s or H100-class hardware. A CPU/RAM combo is generally not viable for a Minecraft server if you want reasonable response times.

You can read more about Ollama here: https://itsfoss.com/ollama/