Public preview · figures are illustrative fixture data, attributed to their source · data to May 2025 · read-only

Models

25 open-weight models and what they need to run

Run locally1 of 25

Needs 22 GB or less at 8K context. Ordered by how many reference systems run it well.Show all models

Coding Run locally Reasoning Vision Agents & tool use Fast models Small models Permissive licenses

Vision ×Clear filters

Gemma 3 27B
Chat model that also understands images. Among the strongest open models at following instructions.
Google DeepMind
- ~20 GB to run · fits a 24 GB graphics card
- Runs well on most reference systems
- Restricted license · Weights are downloadable; the license limits some uses
View model →

Parameters, context length, capability profiles and side-by-side comparison are in the Technical view.