Response times vary by AI model and request complexity. Larger models (Fable 5, Opus, O3, the Pro tiers) take longer than faster models (Haiku, Gemini Flash, the Mini and Nano tiers). Complex tasks like code generation or research may take 30-60 seconds. If an agent seems stuck: check the progress indicator for status updates, try refreshing the page to re-establish the SSE connection, or check the Usage page to see if you've approached your tier's rate limits.