Guanaco vs. MiniMax

Overview

Guanaco is an LLM based off the QLoRA 4-bit finetuning method developed by Tim Dettmers et. al. in the UW NLP group. Guanaco achieves 99% ChatGPT performance on the Vicuna benchmark.

Guanaco is an LLM that uses a finetuning method called LoRA that was developed by Tim Dettmers et. al. in the UW NLP group. With QLoRA, it becomes possible to finetune up to a 65B parameter model on a 48GB GPU without loss of performance relative to a 16-bit model. The Guanaco model family outperforms all previously released models on the Vicuna benchmark. However, given the models are based off of the LLaMA model family, commercial use is not permitted.

Initial release: 2023-05-23

Reference

https://github.com/artidoro/qlora

Further Reading

• https://arxiv.org/abs/2305.14314
• https://twitter.com/Tim_Dettmers/status/1661379354507476994

Overview

MiniMax is an open-weight model family spanning language, reasoning, multimodal, and agentic workloads. MiniMax M3 is the current generation.

MiniMax introduced the open MiniMax-01 series in January 2025 and followed it with the M1 and M2 reasoning and agent models. MiniMax M3, released in June 2026, is a 428B-parameter mixture-of-experts model with about 23B active parameters. It combines native image and video understanding, a one-million-token context window, sparse attention, computer use, coding, and long-horizon agent capabilities.

Initial release: 2025-01-15

Current generation: MiniMax M3

Reference

https://www.minimax.io/blog/minimax-m3

Further Reading

• https://www.minimax.io/news/minimax-01-series-2
• https://www.minimax.io/blog/minimax-m3
• https://huggingface.co/MiniMaxAI/MiniMax-M3