Duck.ai offers a variety of AI models from different providers, each with different strengths to perform different types of tasks. Some models are designed for quick, everyday tasks, while others are built to handle more complex work like coding, analysis, or long-form writing.
Each prompt you send in a chat has an associated cost against your usage limit, which resets after regular intervals. More advanced models use more of your limit per prompt, so choosing the right model for the task at hand can help you get the most out of Duck.ai. For simple questions or casual conversations, start with a low- or extra low-cost model. Save high-cost models for more complex tasks that need them.
Some models also support reasoning, which lets the model spend extra effort working through a problem before answering, and “extended reasoning” mode gives the model even more time and compute to think. A good rule of thumb is to use a model that has a reasoning mode if you want a balance between response speed and accuracy, and enable extended reasoning when the prompt you're asking the model is especially complex or when accuracy matters more than speed. Although models with reasoning can produce better results for complex tasks, they have a higher cost and use more of your usage limit.
Model Comparison
| Model | Limit Cost | Reasoning | Extended Reasoning |
| Claude 4.5 Haiku | Medium | ✔ | |
| Mistral Small 4 | Extra Low | ||
| GPT 5.4 nano | Extra Low | ✔ | |
| GPT 5.4 mini | Low | ✔ | ✔ |
| gpt-oss-120b | Extra Low | ✔ | |
| Gemma 4 31B | Extra Low | ✔ | |
| GPT-5.4* | High | ✔ | ✔ |
| Claude Sonnet 4.6* | High | ✔ | |
| Claude Opus 4.8* | Extra High | ✔ | ✔ |
*Requires a DuckDuckGo subscription
Example Use Cases
Choosing the right model often comes down to matching the complexity of your task to the appropriate cost tier. Here is a general guide to what each tier handles best.
Extra-low-cost models are best for fast, lightweight tasks where volume or speed matters more than depth. This includes quick brainstorming, simple copy edits, turning bullet points into a short social post, writing or paraphrasing short text, creating simple outlines, and handling straightforward technical tasks like writing spreadsheet formulas or answering quick coding questions.
Low-cost models sit in a useful middle ground, covering tasks that need a bit more reasoning without justifying a higher spend. Good examples include working through homework-style problems, polishing a resume, or comparing a few options before making a decision.
Medium-cost models are well suited for the kinds of writing and communication tasks that come up constantly in a workday, such as drafting emails, summarizing meeting notes, and rewriting a document in a clearer tone.
High-cost models handle more demanding work and tend to split between technical depth and writing quality. This tier is appropriate for debugging tricky code, planning complex projects, analyzing detailed documents with tradeoffs and edge cases, synthesizing research, and turning rough ideas into polished first drafts.
Extra-high-cost models are best reserved for the most open-ended and high-stakes tasks, including deep strategy work, long-form writing that requires nuance and precision, and complex analysis that spans many competing constraints.