Open-Weight Models in 2026
Updated
Open-weight models can be relevant when you want to assess model options alongside the systems and constraints in your stack. This catalogue includes entries marked open, and its ordering is by context, highest first. Llama 4 Scout, MiMo-V2.5, and MiMo-V2.5-Pro are among the models shown, alongside additional MiMo, LongCat, and DeepSeek variants. Use the context ordering to navigate the entries, then filter or sort around the model attributes that matter to your workload. The catalogue designation identifies the models included here as open-weight; consider how each candidate fits your intended use before settling on one.
178 on record. Sorted by context; every figure comes from the OpenRouter model catalogue, with the page it was read from on each entry.
Context bars share one log scale, 1K to 2M tokens. Prices are what each provider publishes per million tokens; 'varies' is a router that bills the model it picks.