Output
Output
Awaiting query...
No active inference
Submit a query to begin
Smallest packed context, lowest latency, strictest confidence gate. Context budget ≈ 10% of each model window (max 16,000 input tokens after packing). Logprob gate ≥ 55%.
Awaiting query...
No active inference
Submit a query to begin