Standard vs. Advanced Models FAQs
Understand the differences between using standard and advanced models in Astell
What are Standard Models?
Standard Models are fast, reliable models designed for everyday work. They consume your existing Astell token balance (the same shared pool used for ingestion).
What are Standard Models good for?
Standard Models are ideal for:
- Everyday questions and quick answers
- Summarizing messages and documents
- Draft generation and brainstorming
- Code explanations and debugging assistance
- Meeting note summaries
- Email composition
- General knowledge queries
What does “unlimited” mean for Standard Models?
Standard Models are no longer billed with a separate message quota. Instead, usage is token-based and draws from your monthly token pool.
What are Advanced Models?
Advanced Models are the latest, state-of-the-art reasoning models intended for more complex tasks, typically with longer context windows. They also consume your Astell token balance.
Which plans can use Advanced Models?
- Sapling: not available
- Tree: available
- Grove: available
- Forest (Enterprise): available (custom terms as needed)
How is chat token usage measured?
Astell charges based on what you send in chat and what the agent sends back. Internal/background model usage (for example, memory management and other system operations) is not charged to your token balance.
What are Advanced Models best for?
Advanced Models are best for:
- Complex analysis and research
- Detailed technical documentation
- High-stakes business decisions
- Advanced code generation
- In-depth strategic planning
- Nuanced writing and editing
- Multi-step reasoning tasks
Do context, memory, or long multi-turn conversations cost tokens?
Context and memory features do not have their own separate fee, but each chat turn still consumes tokens based on the message exchange. Tokens are also consumed for Processed Data ingestion (See Native vs. Processed Data).
How do file/document attachments affect token usage in chat?
If a file needs Processed Data handling (for example, OCR, transcription, or media parsing), the related ingestion cost still applies in addition to normal chat usage.
What's the recommended way to choose the right model?
Start with Standard Models first (they cover about 80–90% of typical use cases). Upgrade to Advanced Models when:
- The standard response isn't detailed enough
- You need more nuanced understanding
- The task requires advanced reasoning
- You're refining important work
Can I switch models mid-conversation?
Yes, you can always switch between standard and advanced models at any time when using Astell. To switch models in chat, please see "AI Models".
How do I conserve usage in advanced chat (without sacrificing results)?
The simplest rule: start with Standard Models, then switch to Advanced Models only when you actually need deeper reasoning or higher accuracy. Standard models handle 80%+ of everyday questions well, and the biggest waste usually comes from unnecessary back-and-forth.
- Batch your questions (combine multiple asks into one request).
- Be specific up front so you don't burn extra turns on follow-ups.
- Use one “do-it-all” prompt instead of three separate prompts.
Example: - ❌ Wasteful (3 responses): three separate Advanced responses
- ✅ Efficient (1 responses): one Advanced response: “Analyze this document: summarize the content, identify key risks, and list action items with priorities.”
What happens when I run out of tokens?
- Standard and advanced chat can pause once your available token balance is exhausted
- Processed Data ingestion can also pause
- Usage resumes when tokens reset, pay-as-you-go is enabled, or your plan/token limit is updated