Muse Spark 1.3 on Thinkr.

• Open

  Open holds the models that no provider serves with ZDR, every video model among them. The provider reads the chat and may keep it for a while.

In ZDR and Open, Thinkr's proxy can read the chat in transit and writes none of it to disk.

• Llama 3.3 70BMeta's open-weight model for fast chat in many languages.See the model
• Muse Spark OpenMeta's open-weight Muse Spark, trained for agentic coding.See the model
• Muse Glimmer 30BMeta's open-weight 30B model, distilled from Muse Spark, for agents.See the model