Choosing an AI Model

Every model-enabled feature starts on Auto, so you do not need to understand model names to get a strong result. You can still open the model picker and choose a named model whenever you want more control.

Default model by plan and feature

Reasoning effort describes how long the automatic model is allowed to think, review, and improve its answer or creation. Low prioritizes efficiency, Medium adds more deliberation, and High gives the model the most time for complex work.

Auto always stays on a standard model, so a creation started on Auto costs one AI creation on every plan. Choosing a Frontier model by name is what changes the cost.

The AI Image Playground uses a separate image-model catalog: Nano Banana 2 Lite is its default on Free, Premium, and Laureate. Text reasoning effort does not apply. Voice Mode also uses a separate real-time voice system.

What a model costs

The model you choose decides how much a creation weighs against your weekly AI creation allowance:

• Standard models cost 1 AI creation. This covers every model except the two below, including everything Auto uses.
• GPT 5.6 Sol and Claude Opus 5 cost 3. They are our Frontier models, and their rows in the picker carry a 3× badge so you can see the cost before you choose.

Choosing a Frontier model does not draw on a separate allowance — it simply weighs more against the same one. Every paid plan can choose from the full creation model catalogue, Frontier models included.

Sol and Opus are for creations only. Unlimited AI Chat runs on the standard models, and the model you pick in Chat never changes what a message costs.

A failed creation returns what it reserved, and weekly allowances reset every Monday at 00:00 UTC.

Where you can pick a model

Model selection is available in:

• Home and AI Chat
• AI Slides and document creators
• Deep Research
• AI Video Lectures
• Podcasts
• Flashcards on paid plans
• Worksheets, Study Guides, Mind Maps, and Infographics on paid plans

The picker opens on Auto, Auto (Premium), and Auto (Laureate), shown on every plan so the plan differences are visible before upgrading. Auto rows name the plan tier only — they no longer show the model behind them, so an Auto choice keeps following current routing instead of freezing to one model.

Named models sit below Auto. In creation tools the main menu lists Frontier Models, then Recommended Models, with the rest of the catalogue one level deeper under More Models. AI Chat has no Frontier models, so its menu goes straight from Auto to Recommended Models.

A model that needs a paid plan carries an Upgrade badge. Choosing it opens your plan options rather than failing when you send.

Scholarly remembers the model you chose for each creation tool separately — including Podcasts, Video Lectures, Research, and Slides — and remembers your Chat model separately again, so returning to a tool starts from your last pick.

For Chat and most creation tools, the Recommended Models are GPT 5.6 Terra, Grok 4.6, Claude Sonnet 5, and Gemini 3.1 Pro Preview. Flashcards use a focused recommendation: Gemini 3.8 Flash. Free users can browse these paid choices, and selecting one opens your plan options.

Models grouped under Frontier Models are the most capable, higher-compute choices: GPT 5.6 Sol and Claude Opus 5. Any paid plan can pick them, and each one costs 3 AI creations instead of 1.

Named models

The exact catalog changes as models are released or retired. Current choices across features include GPT 5.6 Sol, Terra, and Luna; Gemini 2.5 Flash, 3 Flash, 3.5 Flash Lite, 3.8 Flash, and 3.1 Pro Preview; Grok 4.6; Deepseek Flash; Mistral Small, Medium, and Large; and Claude Haiku 4.5, Sonnet 5, and Opus 5.

Only GPT 5.6 Sol and Claude Opus 5 cost 3 AI creations. Every other named choice costs 1, the same as Auto.

Thinking in Chat

Some named Chat models offer a Thinking control. Turning it on lets the model reason before answering and shows a progress summary that collapses into a completed thought block. Other models always reason and do not show the control.

Model data handling and privacy

Scholarly never sells your content or uses your prompts and sources to train AI models. Your material is sent only to the provider needed to produce your requested result. See the Privacy Policy for full details.

Tips

• If you are unsure, leave the picker on Auto.
• Try another named model when you want a different writing style or a second take.
• Save GPT 5.6 Sol and Claude Opus 5 for the work worth 3 AI creations; everything else costs 1.
• Two other rules still apply whatever you pick: failed requests return their reservation, and a new flashcard deck with fewer than 15 cards does not count.