Skip to main content
Quill AI supports models from OpenAI, Google, and Anthropic. Each model has different strengths, costs, and speeds. This guide helps you choose the right one for your task.

Understanding Model Properties

Reasoning ability:
  • Models with reasoning think through problems step-by-step
  • Better for complex rewrites, maintaining consistency, and story logic
  • Slower but more thorough
Context window (tokens):
  • How much text the model can process at once
  • Larger windows = can handle more of your script at once
  • Important for long scenes or multi-scene analysis
Speed (★ rating):
  • How quickly the model responds
  • ★★★★ = Fast (seconds)
  • ★★ = Slower (10-30 seconds)
Cost (credits per request):
  • Each request deducts credits from your account
  • Lower cost = more requests per month
  • Higher cost often = better quality

Available Models

OpenAI Models

ModelCostSpeedReasoningContextBest For
GPT-4.1 mini0.25★★★★No1M tokensQuick edits, dialogue polish, everyday tasks. Fastest and cheapest.
GPT-4o Mini0.5★★★No200K tokensBudget-friendly for routine revisions and brainstorming.
GPT-4o1★★★No128K tokensBalanced quality and speed for general rewriting.
o4 mini1★★★Yes (medium)200K tokensStructured reasoning for scene restructuring and plot logic.
o3-mini1★★★Yes (high)200K tokensStrong reasoning for complex multi-scene consistency.
GPT-4.11★★★No1M tokensVery large context for analyzing full acts or long sequences.

Google Models

ModelCostSpeedReasoningContextBest For
Gemini 2.0 flash0.5★★No1M tokensBudget option with huge context for long-form analysis.
Gemini 2.5 flash1★★Yes (high)1M tokensHigh reasoning with massive context for complex story work.

Anthropic Models (Creative & Nuanced)

ModelCostSpeedReasoningContextBest For
Haiku 4.50.25★★★No200K tokensFast, creative dialogue and character voice. Often more natural prose.
Sonnet 4.51★★Yes (high)200K tokensMost creative and nuanced. Excellent for character depth and subtext. Pro plan only.
Anthropic models (Claude) tend to produce more creative, nuanced writing with stronger character voice. Try Haiku 4.5 or Sonnet 4.5 if other models feel too generic.

Choosing the Right Model

For Quick Tasks (Speed Priority)

Best: GPT-4.1 mini, Haiku 4.5
  • Fast responses
  • Low cost (0.25 credits)
  • Perfect for: Dialogue tweaks, quick brainstorming, line edits

For Budget Writing (Cost Priority)

Best: GPT-4.1 mini (0.25), Haiku 4.5 (0.25), GPT-4o Mini (0.5), Gemini 2.0 flash (0.5)
  • Maximize your free 100 credits
  • ~200-400 requests per month
  • Perfect for: Daily writing assistance, exploratory brainstorming

For Creative Writing (Quality Priority)

Best: Sonnet 4.5 ⭐ (Pro only), Haiku 4.5, GPT-4.1
  • More natural, nuanced prose
  • Better character voice
  • Perfect for: Dialogue refinement, character development, subtext

For Complex Story Problems (Reasoning Priority)

Best: Sonnet 4.5 ⭐, o3-mini, Gemini 2.5 flash, o4 mini
  • Step-by-step reasoning
  • Better consistency across scenes
  • Perfect for: Plot logic, multi-scene rewrites, structural issues

For Long Scripts (Context Priority)

Best: GPT-4.1 (1M tokens), Gemini models (1M tokens)
  • Can process entire acts at once
  • Better for continuity checks
  • Perfect for: Analyzing full sequences, maintaining consistency

Switch Models

1

Open Quill AI

Press Cmd/Ctrl+L to open the Quill panel
2

Click model selector

Click the model name at the top of the panel (shows current model)
3

Choose new model

Browse by provider (OpenAI, Google, Anthropic) and select a model
4

Confirm

The new model name appears in the selector. Your next message uses this model.
You can switch models mid-conversation. The change applies from your next message onward without affecting earlier responses.

Model Recommendations by Task

Dialogue refinement: Haiku 4.5, Sonnet 4.5, GPT-4.1 mini Plot brainstorming: o3-mini, Sonnet 4.5, GPT-4o Scene structure: o4 mini, Gemini 2.5 flash, o3-mini Character development: Sonnet 4.5 ⭐, Haiku 4.5, GPT-4o Quick edits: GPT-4.1 mini, Haiku 4.5 Full script analysis: GPT-4.1, Gemini 2.5 flash (large context) Budget-friendly daily use: GPT-4.1 mini, Haiku 4.5, GPT-4o Mini

Frequently Asked Questions

There’s no single “best” model—it depends on your task. Sonnet 4.5 (Pro only) is the most creative and nuanced, but costs 1 credit per request. For everyday use, GPT-4.1 mini or Haiku 4.5 (0.25 credits) offer the best value.
Anthropic’s Claude models (Haiku and Sonnet) tend to produce more natural, literary prose with better character voice and subtext. They’re particularly good at dialogue and maintaining consistent character personalities.
No. Reasoning models (o4 mini, o3-mini, Sonnet 4.5, Gemini 2.5 flash) are slower and better for complex problems. For simple edits or brainstorming, regular models are faster and just as effective.
No. Sonnet 4.5 requires a Pro plan subscription. All other models are available on both free and Pro plans.
Models with reasoning capabilities or larger context windows take longer to process. Speed ratings (★) help you choose based on how quickly you need a response.
Not always. Cost reflects computational requirements, not always output quality for your specific task. A 0.25 credit model might be perfect for quick dialogue tweaks, while a 1 credit model shines in complex story logic.

Next Steps