Strengths
- Apache 2.0 with no revenue cap, more permissive than the custom licenses of Kimi K3 or the open Qwen3.8-Max weights.
- Officially supports 48 languages, including Slovak and Czech.
- Cohere says it runs on one B200 or two H100 cards (W4A4 quantization) with negligible quality loss. That is fairly modest hardware for a model this size.
- One model covers images, agents, reasoning and translation.
Weaknesses
- Its context window is 128K tokens, far less than the one million of most competitors.
- Cohere does not publish a per-token price and mainly sells the model through enterprise contracts.
- The model card publishes no benchmark results, so you have to test the quality yourself.
Best for
- answers over company documents and agents on your own servers
- business texts and translation in Slovak and Czech
- regulated industries that want an open model from outside China
Not ideal for
- very long documents over 128K tokens
- individual consumers, as Cohere has no chat app subscription
How to use it in Europe
You can download the weights from Hugging Face under Apache 2.0 and run them on your own server in Europe, so the data never leaves your company. Cohere also offers an API, dedicated Model Vault instances and on-premise deployments, but the price is agreed with sales. A trial API key is free, though not meant for production or commercial use. Cohere offers no chat app subscription for individual consumers.
Similar models from other vendors
- Claude Sonnet 5.5 (Anthropic): Claude's balanced model for everyday work and coding. It is faster than Opus 5.5 and costs half as much.
- Gemini 3.8 Flash (Google): Google's fast, lower-cost model for everyday work, coding and document analysis. You can try it in the API for free.
- Muse Spark 1.3 (Meta): Meta's balanced model with a 1-million-token window and a fair price. The Muse Spark line also powers the Meta AI assistant.
- Muse Spark Open (Meta): Meta's large open model that a company can run on its own servers, so data never leaves its infrastructure.
- Llama 4 Maverick (Meta): The larger model in the newest Llama generation, from April 2025. Cheap through providers, but its license excludes people and companies in the EU.
- Mistral Medium 3.5 (Mistral AI): Mistral's mid-size open-weight model and the default in the Vibe app. Good for everyday work and coding with data kept in the EU.