Our model is bigger and more capable than gpt OSS and can run at full context at 40 tokens / s.

We are rolling out to Mac to start with plans to release windows and Linux within 3 months.