Anthropic Claude Opus and Sonnet Tiers Quiz

Answers and explanations appear as you take the quiz.

1. What naming theme do Haiku, Sonnet, and Opus draw from?
   • A mix of literary and musical terms
   • Greek letters only
   • Planets
   • Colors
2. Within Anthropic’s Haiku, Sonnet, and Opus tiers, how is Opus positioned?
   • It is the fastest option in every model family
   • It is positioned for the most demanding work among Haiku, Sonnet, and Opus
   • It is a speech-only model
   • It is Anthropic’s only model family
3. What approach did Anthropic’s 2022 Constitutional AI paper describe for training a harmless assistant?
   • Training only on image labels
   • Removing the model’s language data
   • Using written principles to guide self-critique and AI-generated feedback
   • Using no evaluation or feedback at all
4. How does Anthropic position Sonnet among the Claude tiers?
   • As a model for image generation only
   • As a retired safety policy
   • As a slower, most expensive tier in every release
   • As a balance of speed and intelligence
5. What does Claude’s extended-thinking mode do when supported?
   • Spends additional computation reasoning through a complex task before answering
   • Responds instantly without computation
   • Turns off safety features
   • Deletes the conversation history
6. Which Claude tier does Anthropic describe as its fastest option?
   • Opus
   • Haiku
   • Sonnet
   • Fable
7. What follows the self-critique stage in the Constitutional AI paper’s AI-feedback approach?
   • The model stops after critique with no further training
   • The tokenizer is replaced with an image encoder
   • AI-generated preference comparisons train a reward model used for reinforcement learning
   • The model is trained only with manually ranked human comparisons
8. Which non-text input modality can Claude models support according to Anthropic’s model documentation?
   • Lidar point clouds for autonomous driving
   • MIDI files as native control commands
   • Robot joint torques
   • Images