Browse 5,696+ open-weight AI models by category
Models that accept image input via URL or base64. Can analyze screenshots, describe images, and perform OCR.
Vision models that excel at extracting text from images, scanned documents, and screenshots.
Instruction-tuned models optimized for conversational use, general Q&A, and assistant tasks.
Models specialized in code generation, debugging, and software engineering tasks.
Models trained for complex reasoning, logical deduction, and step-by-step problem solving.
Models specialized in mathematical reasoning, calculations, and numerical problem solving.
Models fine-tuned for creative writing, storytelling, and roleplay scenarios.
Models with strong multilingual capabilities, supporting translation and non-English tasks.
Models fine-tuned for medical terminology, clinical documentation, and healthcare tasks.
Models trained on legal documents, contracts, and regulatory compliance.
Models trained for cybersecurity analysis, threat detection, and vulnerability assessment.
Models used for RLHF reward scoring, preference ranking, and evaluation of other models.
Models that generate vector embeddings for semantic search, retrieval, and re-ranking.
Compact models with 3 billion parameters or fewer. Fast, low-cost, suitable for simple tasks.
High-capacity models with 30 billion parameters or more. Best quality but slower and more expensive.
Base models and models not matching other categories. May require additional prompting or fine-tuning.
Or open a dedicated playground for any model from the catalog.