Compare LLM Models Side by Side

Select models to compare

Choose up to 4 models for a side-by-side comparison of benchmarks, VRAM requirements, and real-world performance.

Explore any model or compare side by side

Search our database of 290+ models, or paste any HuggingFace model ID to get instant VRAM estimates — even for models we don't track yet. Select one model for a detail view, or add more to compare side by side.

From our database: Llama 3.1 8B, Qwen 2.5 7B, Phi-3 Mini · From HuggingFace: meta-llama/Llama-3.1-8B, mistralai/Mistral-7B-v0.3