Capability indicators
When you hover over a model in the assistant selector, four indicators appear: Speed — how fast the model responds, on a scale from 1 to 5. Faster models are better for quick tasks like summarizing a paragraph or translating a sentence. Slower models tend to be more thorough. Intelligence — how nuanced and accurate the model’s reasoning is, on a scale from 1 to 5. Higher intelligence means the model is better at complex analysis, multi-step reasoning, and catching subtle distinctions. The trade-off is usually speed. Capacity — how much text the model can process in a single conversation, shown in approximate words or pages. This matters when you’re working with long documents or asking questions that require the AI to consider a lot of material at once. Knowledge cutoff — the date up to which the model was trained. The model doesn’t know about events, legislation, or case law published after this date. If you need current information, combine the model with web search.Reasoning
All models in Saga are configured for high reasoning by default, so you do not need to adjust a separate reasoning setting. Choose the model based on the type of work you want to do, such as speed, intelligence, capacity, or access to current information.Auto
If you’re not sure which model to use, select Auto from the assistant selector. Saga will analyze your message and automatically pick the best model from the ones your organization has enabled — balancing cost, speed, and intelligence. Auto works well for everyday tasks. It routes simple questions to fast, lightweight models and sends complex analysis to more capable ones. You don’t need to think about which model fits — Saga decides based on what you’re asking. When Auto selects a model, you’ll see an “Answered by” label on the response showing which model was used. This way you always know what’s behind the answer, even when you didn’t choose it yourself. You can set Auto as your default in Settings > Personalization, or your admin can set it as the workspace default. It’s also available as a project default.Choosing the right model
If you prefer to pick a model yourself, there’s no single best choice — it depends on what you’re doing.- Quick tasks (summarize, translate, draft a short response): pick a fast model for a quicker response.
- Complex analysis (multi-party contract review, regulatory comparison, risk assessment): pick an intelligent model. The extra processing time is worth the more nuanced output.
- Large document context (long contracts, extensive due diligence sets): check capacity. A model with higher capacity can hold more of your material in context at once.
- Current information (recent legislation, recent case law): check the knowledge cutoff, and enable web search if needed.