Selecting the Right AI Model

Magicdoor lets you use multiple AI providers in one interface. The practical benefit is that you can choose the model that fits the task instead of forcing one provider to do everything.

Strong starting point

Claude Sonnet 5 is a strong general-purpose starting point for writing, analysis, code, and most everyday work.

Current chat lineup

• Claude Opus 4.8: Highest-cost Claude option for harder reasoning work
• Claude Sonnet 5: Strong default for writing, analysis, and code
• Claude Haiku 4.5: Lower-cost Anthropic option
• GPT-5.5: OpenAI flagship option for broad general work
• GPT-5.4 Mini: Lower-cost OpenAI option for summaries, extraction, and drafts
• Gemini 3.1 Pro: Strong for long documents and multimodal tasks
• Gemini 3 Flash: Fast, lower-cost Google option
• Grok 4.3: Another general-purpose model in the lineup
• sonar-reasoning-pro: Best when you need current information with sources
• sonar-deep-research: Heavier web research workflow
• GLM-5.1: Lower-cost reasoning-oriented model
• DeepSeek V4 Pro: Long-context reasoning model with competitive pricing
• MiniMax M3: Lowest-cost chat option for everyday tasks
• Kimi K2.7 Code: Multimodal long-horizon model for agentic work

Simple selection guide

• Use Claude Sonnet 5 when you want a strong default.
• Use GPT-5.5 when you want the OpenAI option for general work.
• Use GPT-5.4 Mini, Claude Haiku 4.5, or Gemini 3 Flash when cost matters.
• Use Gemini 3.1 Pro when you are working with long documents or multimodal input.
• Use sonar-reasoning-pro or sonar-deep-research when the task depends on current web information.
• Use Claude Opus 4.8 when the reasoning quality matters more than cost.
• Use GLM-5.1, DeepSeek V4 Pro, MiniMax M3, or Kimi K2.7 Code when you want a Together AI model with a different cost/performance profile.

Why switching models matters

One model is not best at everything. A common workflow is:

• research with sonar-reasoning-pro
• synthesize with Claude Sonnet 5 or GPT-5.5
• drop to GPT-5.4 Mini or Gemini 3 Flash for cheaper follow-up work

Magicdoor lets you switch models during a conversation so you can change tools without losing context.

Managing cost

The biggest cost drivers are:

• how long the conversation has become
• which model you are using
• whether the workflow includes image generation or web-connected research

Long chats cost more because the model has to process more prior context. Premium models also cost more than Mini, Flash, or Haiku-style options. If you want concrete examples, see cost per model.

Tips

• start a new chat when you switch topics
• use lower-cost models for drafts and summaries
• save premium models for high-value work
• switch to sonar-reasoning-pro only when you actually need live web results

FAQ

Which model should I start with?

Claude Sonnet 5 works well for many tasks, including writing, analysis, coding, and general conversation. Switch when your needs change.

What's the cheapest model?

GPT-5.4 Mini, Claude Haiku 4.5, Gemini 3 Flash, and GLM-5.1 are the budget options. Gemini 3 Flash at $0.50/$3 per 1M tokens is the cheapest for input/output.

When should I use sonar-reasoning-pro or sonar-deep-research?

Use sonar-reasoning-pro for quick web-backed answers with citations. Use sonar-deep-research for thorough investigation of a topic—it does deeper searching and synthesis.

Can I switch models mid-conversation?

Yes. Magicdoor preserves your conversation context when you switch models. This lets you use different models for different parts of a task.