LLMs, image, video, audio, 3D, avatar, and platform tools in one place. Narrow the list by output type, prompt support, context window, parameters, license, VRAM, pricing, and more — then compare the shortlist side by side.
Choose what you want to build and what matters most. The catalog is ranked instantly and the shortlist can be re-sorted by quality, speed, price or context — no forms, no 400 spec sheets.
aibender is an independent, filterable catalog of 423 AI models and tools, assembled from public vendor pages, model cards, leaderboards and community hardware measurements. Instead of a plain list of names, every record carries the details that decide whether a model fits a real project: context window, parameter count, license, commercial-use terms, VRAM and RAM requirements, pricing, input and output formats, and hard limits such as clip length or outputs per request.
Use the filters to narrow the catalog to a shortlist, then open the compare view to put those models next to each other. Each entry also links to a dedicated page with full specifications, strengths, limits and alternatives from the same vendor.
Separate large language models for chat, reasoning and code from image generators, video models, audio and music models, 3D asset tools, avatar tools and hosted platforms, then add input modalities such as image, audio, video or PDF.
Set a minimum context window from 32K up to 2M tokens, or a minimum output budget from 8K to 1M tokens, so long documents and long generations stay inside the model's published limit.
Compare open-source, open-weights, proprietary and free-tier options, and see whether commercial use is allowed, conditional or unstated before you build on a model.
Filter by minimum VRAM from 8 GB to 192 GB to find models that fit your GPU, or switch to API-only tools when you would rather pay per token, second or image instead of buying hardware.
Start with the output type you need, because a text model, an image model, a video model and an audio or 3D tool solve different problems. Next apply the constraints that cannot be negotiated: license and commercial terms, available VRAM or budget, and the context or clip length the task requires. Only then compare quality signals such as benchmark results, supported languages, resolution, audio support and ecosystem maturity. When two candidates are close, the compare view shows the differences field by field.
The catalog is refreshed as vendors publish new versions, prices and limits. Records that could only be confirmed through secondary sources are marked with a lower confidence level, so you can tell verified specifications from early or unconfirmed ones. Pricing, availability and limits change often: treat this catalog as a research starting point and confirm the final figures on the vendor's own page before you commit.
Explore the full AI model directory, the AI coding editors and agents directory, the browser-based data analysis workspace, or the blockchain AI agents guide.
