Arcee AI: Trinity Mini (free)

Arcee AI: Trinity Mini is a powerful and free-to-use 26B-parameter sparse mixture-of-experts (MoE) language model, featuring 128 experts with 8 active per token. This advanced architecture is specifically engineered for highly efficient reasoning, making it exceptional for processing and understanding long contexts up to 131,000 tokens. Designed for versatility, Trinity Mini offers robust function calling capabilities and supports complex multi-step agent workflows, enabling sophisticated AI applications. It's best suited for chat-based interactions, providing intelligent and coherent responses. With a maximum output of 4,000 tokens and a free access tier, it's an accessible yet powerful tool for developers and researchers. Note that this model does not support image generation.