Skip to main content

​
Overview

Auto Select is the default and works well for most interviews: it starts answering fast and reasons well. Pick another model when you want deeper reasoning for hard technical rounds, or the fastest possible start for rapid back-and-forth.
There’s no single “perfect” model. Test a few in the Playground with real interview questions before your interview.

​
How to Read the Model Picker

Every model in the picker shows two things:
  • Intelligence - how deeply the model reasons: Basic, Medium, High, or Highest.
  • First word - how long you wait before the answer starts to appear: Fastest, Fast, Moderate, or Slower.
Smarter models reason more deeply but take longer to start answering. In a live interview, a fast start often matters more than the last bit of depth. Click a column header to sort by name, intelligence, or speed.

​
Models by Plan

We keep adding the latest AI models as they become available. The model list is subject to change.

​
Included on Every Plan

​
Pro Plan

​
Premium Model Credits

These models use premium model credits, one per session.

​
Quick Recommendations by Interview Type

​
Behavioral and Conversational Interviews

Top Picks:
  • Auto Select - Fast start with strong, natural answers
  • GPT-6 Luna - Quick, well-rounded responses on every plan
  • Claude Haiku 5.5 - Thoughtful answers without a long wait (Pro)
Why these work: Behavioral rounds move quickly and need natural-sounding answers. A fast first word keeps you in the flow of the conversation.

​
Technical Coding Interviews and Online Assessments

Top Picks:
  • Claude Sonnet 5.5 - Top-tier reasoning for hard problems and clear code explanations (premium credits)
  • GPT-6.1 Sol - Deep reasoning for complex algorithms (premium credits)
  • GLM-5.3 - High intelligence with a fast start (Pro)
  • GLM-5.3 Flash - A capable choice on every plan
Why these work: Coding questions reward correct logic and step-by-step explanation, so intelligence matters more than in other rounds. Alternative: For rapid-fire or timed questions, Auto Select or GPT-6 Luna start answering sooner.

​
System Design Interviews

Top Picks:
  • Claude Sonnet 5.5 - Strong architectural thinking and trade-off analysis (premium credits)
  • GPT-6.1 Sol - Deep reasoning for scalability discussions (premium credits)
  • Claude Haiku 5.5 or GLM-5.3 - High intelligence with a fast start (Pro)
Why these work: System design requires high-level thinking and clear explanation of trade-offs, where the smartest models show the biggest difference. Alternative: Auto Select handles general system design discussions well.

​
Data Science, ML, and Case Interviews

Top Picks:
  • Claude Sonnet 5.5 or GPT-6 Sol - Careful, structured reasoning (premium credits)
  • Gemini 3.8 Flash - High intelligence when a slightly longer wait is fine (Pro)
  • Auto Select - A solid default on every plan
Why these work: These rounds need precise reasoning and structured thinking, but usually leave you a little more time to answer.

​
Interviews in Other Languages

All models answer in the interview language you set. For interviews in Chinese, GLM-5.3, DeepSeek V4.1 Flash, and Qwen3.8 Max are good ones to try. Test your language in the Playground first. Tip: Check the change interview language guide to set your preferred language.

​
How to Find Your Perfect Model

1

Start with the recommended model for your interview type

Use the recommendations above as your starting point based on your upcoming interview.
2

Test in Playground with real questions

Go to the Playground and test with actual questions you expect in your interview.
3

Compare 2-3 models side-by-side

Try the same question with different models to compare response style, speed, and quality.
4

Consider your personal preference

Some people prefer concise answers, others want detailed explanations. Choose what feels natural to you.
5

Test different response styles

Adjust the response style to fine-tune the output format. For advanced customization, try creating custom response styles with personalized prompts.
6

Apply and practice

Once you find your preferred setup, apply the settings and run a mock interview session to validate.

​
Common Questions

Pick it when you launch a session, or set your default model in Settings. Choose and test it in the Playground before the interview starts.
Not necessarily. Models like Claude Sonnet 5.5 excel at complex reasoning, but they take longer to start answering. For quick back-and-forth, Auto Select or GPT-6 Luna are often more practical. It depends on your interview type and pace.
Yes, if you like. Use a higher-intelligence model for coding and system design rounds, and a faster one for behavioral rounds. Test each scenario in the Playground first.
Start with Auto Select, or a high-intelligence model like Claude Sonnet 5.5 or GLM-5.3. These handle both coding and high-level reasoning well.
Ask yourself:
  • Are responses coming at the right speed for my interview pace?
  • Do explanations feel natural and easy to understand?
  • Is the technical accuracy high?
  • Does the tone match what I’d say in my own words?
If you answer “yes” to most of these, you’ve found your model.
Yes. Auto Select and GPT-6 Luna are fast and capable, and they’re a good fit when speed matters more than deep reasoning. Test them first to make sure the quality meets your needs.
We replaced the previous generation of models with newer ones. If your saved model is no longer offered, your sessions use Auto Select until you pick another.