# ⚠️ Agent failed before reply: Model context window too small (4096 tokens). Minimum is 16000. i've tried many different api's from nvidia, nothing works