GGUF Explorer

Tiniest GGUF model, instant loading, good for testing

Very small, surprisingly capable for its size

Tiny Qwen with great quality for its size

Nomic Embed Text v1.5, high-quality text embeddings for semantic search

MxBai Embed Large v1, top-quality English embeddings

BAAI BGE Small, lightweight English embeddings, great for on-device semantic search

Jina Reranker v2 multilingual, high-quality cross-encoder reranking for search

BAAI BGE Reranker v2 M3, multilingual cross-encoder reranking with long context

Classic tiny Llama, fast and reliable

Llama 3.2 1B, great for simple tasks with huge context

Qwen 2.5 1.5B, strong multilingual support

Qwen3 0.6B, fast multilingual reasoning with hybrid thinking

Code-specialized Qwen 2.5, great for programming tasks

Largest SmolLM2, excellent efficiency per parameter

Microsoft Phi-3.5, excellent reasoning and coding

Google Gemma 2, strong instruction following

Llama 3.2 3B, excellent quality with huge context

Qwen 2.5 3B, high quality multilingual generation

Qwen3 1.7B, strong multilingual reasoning with hybrid thinking

DeepSeek R1 distilled to Qwen 1.5B, reasoning/thinking model

Microsoft Phi-4, strong reasoning and coding

Qwen 2.5 Coder 7B, best code generation quality

Mistral 7B, strong general performance

Llama 3.1 8B, best quality for capable devices

Qwen3 4B, excellent multilingual reasoning and code generation

DeepSeek R1 distilled to Qwen 7B, strong reasoning/thinking model

Google Gemma 4 E2B, 2.3B effective params (PLE), strong multilingual + reasoning + vision

Google Gemma 4 E4B, ~4B effective params (PLE), top quality at its size class + vision

Hcompany Holo2 4B UI-grounding VLM, vision + text. Best for browser-agent / GUI navigation tasks.

Hcompany Holo2 8B premium UI-grounding VLM, vision + text. Highest-quality grounding for capable devices.