Models
Models is a secondary category: the durable search is often “open alternative to a closed API,” while the new-term search is the model’s own name. Pages here cover LLMs (Kolibri-1), embeddings (EmbeddingGemma 2, BGE-M3), decision models (Cloudflare Clef), and image generation (Qwen-Image-2.1-Turbo).
We do not invent VRAM figures or leaderboard scores. If a hardware row says “not yet verified,” it is because the seed data and verification notes did not include it. License strings are copied from the Hugging Face card or GitHub LICENSE as of 2026-10-13; “other” means the card does not map to a simple SPDX id.
Start from the hubs: closed LLM APIs, embedding APIs, classification APIs, and Midjourney-class generation.
Also see: Image & Design · Video & Audio · Writing & Office · Dev & Coding · AI Chat & Agents
Alternatives in this category
Latest tools
No tools tagged in this category yet.
Models
- BGE-M3 — BAAI multilingual multi-function embedding model (MIT).
- Cloudflare Clef — 27B multimodal decision model that returns option probabilities in one forward pass.
- EmbeddingGemma 2 — Google DeepMind open multimodal embedding model (~0.74B, Apache-2.0).
- Kolibri-1 — Aleph Alpha German/English MoE reasoning model with tool calling and 1M context.
- Qwen-Image-2.1-Turbo — Accelerated Qwen-Image-2.1 checkpoint for text-to-image and editing (weights released).