Smaug Flash announced as an open-weights model tuned for personal agents
The announcement says the fine-tuned model is available through the RouteLLM API and claims performance on par with DeepSeek Flash.
TLDR
The September 12 announcement pitches Smaug Flash as an open-weights fine-tune for personal agents. It lists RouteLLM API pricing as "$0.10 M input" and "$0.40 M output," while claiming DeepSeek Flash-level performance at a lower cost.
Combined views
—
2 Sources, first seen ago
