Model Selection Strategy

You are mid-feature, the agent just produced a solid implementation plan, and now you need it to write the actual code. Do you stick with the default model or switch? You have heard Opus 4.7 is “the best” but it is also the most expensive. Sonnet 4.6 is cheaper but you are not sure if it can handle the complexity. Gemini 3 Pro has a massive context window but you have never tried it. Composer 2 is the Cursor 3.0 default in Auto mode. Meanwhile, the model picker shows nine options and zero guidance. This article gives you the decision framework so model selection becomes a two-second choice, not a five-minute debate.

What You’ll Walk Away With

• A clear default model recommendation (and why it is the default)
• A decision tree for when to switch models based on task type, not guesswork
• Cost-per-task estimates so you can budget your monthly usage
• Practical prompting adjustments for each model’s strengths
• The keyboard shortcut to switch models instantly without leaving your flow

The Quick Reference

Model	Input / Output Cost	Default Context	Max Context	Best For
Composer 2	$1.25 / $10 per 1M tokens	200k	—	Cursor 3.x default in Auto mode; speed-critical iteration
Claude Opus 4.7	$5 / $25 per 1M tokens	1M	1M	Top-tier agentic coding, complex reasoning, xhigh effort level
Claude Sonnet 4.6	$3 / $15 per 1M tokens	1M	1M	Budget-conscious daily work, extended thinking
Claude Haiku 4.5	$1 / $5 per 1M tokens	200k	200k	Fastest Claude, near-frontier quality on focused tasks
Gemini 3 Pro	$2 / $12 per 1M tokens	200k	1M	Extreme context, multimodal (image/diagram analysis)
GPT-5.5	$5 / $30 per 1M tokens	1M	1M	OpenAI’s latest frontier — strong agentic coding, computer use, research
Grok Code	$0.20 / $1.50 per 1M tokens	256k	—	Very budget-friendly simple tasks

Note

Cursor’s pricing is usage-based at API rates. Your Pro plan ($20/month) includes a budget of usage that gets consumed based on which models you use. Opus 4.7 burns through it faster than Sonnet 4.6 or Composer 2, but the quality difference on complex tasks is substantial. Composer 2 is the default when Auto mode is on — manual model selection overrides it.

Composer 2: The Cursor 3.x Default

Since Cursor 3.0 (April 2, 2026), Composer 2 is the default model in Auto mode. It’s Anysphere’s own coding-first model, trained on top of the Kimi K2.5 base with continued pretraining and a 4x RL run. CursorBench 61.3 (up from 44.2 on Composer 1.5), 200+ tokens/sec on custom GPU kernels, and typical agent turns under 30 seconds.

When to Stick with Composer 2

• Rapid iteration and style tweaks where you’ll refine 5-10 times anyway
• Well-scoped refactors inside a single file or module
• “Quick question” inline edits while working on a larger plan
• Running multiple agents in parallel in the Agents Window — Composer 2’s speed keeps all panes responsive

When to Switch Away

• Task touches cross-module architecture or security-sensitive code — Opus 4.7
• You need the model to reason about a 500k-line codebase — Gemini 3 Pro or Opus 4.7 in 1M mode
• The current model keeps getting a detail wrong — a fresh model often breaks the loop

Claude Opus 4.7: Top-Tier Coding and Reasoning

Claude Opus 4.7 is Anthropic’s flagship generally-available model (released April 16, 2026). It has the highest SWE-Bench scores, the best agentic performance (it handles multi-step tool use reliably), and is the first Claude model with high-resolution image support (2576px / 3.75MP) for screenshot/artifact analysis. In Cursor it’s available on the model picker and in Auto High / Max modes.

When to Use Opus 4.7

• Complex multi-file refactoring
• Architecture design and system planning
• Security audits and code review
• Test generation for nuanced business logic
• Computer-use / screenshot workflows (2576px resolution with 1:1 pixel mapping)
• Any task where getting it right the first time saves more money than the model costs

Prompting Tips for Opus 4.7

Opus 4.7 ships with xhigh effort level as a new top setting. It also follows instructions more literally — if your Opus 4.6 prompts had scaffolding like “double-check the slide layout before returning”, try removing it and re-baseline. Response length now calibrates to perceived task complexity, so expect leaner output on simple asks.

Copy-paste prompt for Opus 4.7 architecture work

Analyze the authentication system across all files in src/auth/ and src/middleware/.
Identify security vulnerabilities and architectural issues.
Propose a refactoring plan that addresses each issue.
Before implementing, explain your reasoning and ask if I want to adjust the approach.

Claude Sonnet 4.6: The Budget Workhorse

Sonnet 4.6 costs 40% less than Opus 4.7 on input tokens and 40% less on output tokens, with a full 1M context window. For straightforward tasks — writing a utility function, adding a field to a form, generating standard CRUD endpoints — the quality difference is minimal.

When to Switch to Sonnet 4.6

• Routine coding where the pattern is well-established
• Tasks where you will review and iterate regardless
• When your monthly usage budget is running low
• Long conversations where you need 1M context but want to manage cost

The Practical Test

If you can describe the task in one sentence and the expected output is predictable, Sonnet 4.6 will handle it. If the task requires weighing tradeoffs or understanding subtle architectural implications, stick with Opus 4.7.

Copy-paste prompt for Sonnet 4.6 utility work

Create a TypeScript utility function that debounces async functions.
It should support cancellation and return the result of the last invocation.
Include comprehensive tests using Vitest.

Claude Haiku 4.5: Fast, Cheap, Near-Frontier

Haiku 4.5 (claude-haiku-4-5) is the fastest Claude at $1/$5 per MTok with 200k context. It’s meaningfully better than anything in its tier for focused tasks: short refactors, code explanation, lint-style feedback. In the Agents Window, drop it into one pane as a “fast reviewer” that inspects Opus 4.7’s output while Opus keeps iterating.

Gemini 3 Pro: The Context Specialist

Gemini 3 Pro’s headline feature is its 1M token context window (accessible via Max mode). When you need the AI to understand your entire codebase at once — not just the files you manually reference — Gemini 3 Pro is the model to reach for.

When to Switch to Gemini 3 Pro

• Analyzing large codebases (50k+ lines) where cross-module understanding matters
• Working with images — paste a screenshot of a UI bug or a Figma design directly into chat
• Reviewing architectural diagrams or documentation that includes visual elements
• Tasks where context volume matters more than reasoning depth

Multimodal Advantage

Gemini 3 Pro handles images natively. Drag a screenshot into the Cursor chat and ask it to reproduce the layout, identify the visual bug, or implement a design from a mockup.

Copy-paste prompt for Gemini 3 Pro visual work

I've attached a screenshot of our current dashboard. The spacing between the
metric cards is inconsistent and the chart legend overlaps on mobile.
Fix the responsive layout in src/components/Dashboard.tsx to match the
spacing from our design system (8px base grid).

GPT-5.5 and Gemini 3 Pro: The Alternative Frontier

GPT-5.5 (released April 23, 2026) is OpenAI’s latest frontier model, available in Cursor’s model picker. It excels at agentic coding, computer use, knowledge work, and research workflows. At $5/$30 per MTok with 1M context, it competes directly with Opus 4.7. Use it when you want a different “perspective” — sometimes switching model families unsticks a problem that one family keeps getting wrong. /best-of-n (new in Cursor 3.0) runs the same task across Composer 2, Opus 4.7, and GPT-5.5 in parallel worktrees and compares outcomes.

Gemini 3 Pro remains the headline choice for multimodal work. Drag a screenshot into the Agents Window or Design Mode and ask it to reproduce the layout, identify a visual bug, or implement from a mockup.

The Decision Tree

When a new task comes in, run through this:

1. Is Auto mode on and the task lightweight?

   Yes: Composer 2 handles it in seconds. Let Cursor do its thing.

2. Is this a complex, multi-file task or architectural decision?

   Yes: Use Claude Opus 4.7 (xhigh effort for deepest reasoning). Or try /best-of-n to compare against GPT-5.5 and Composer 2.

3. Is this a simple, well-defined task with predictable output?

   Yes: Use Claude Sonnet 4.6 or stick with Composer 2 (if speed matters). Claude Haiku 4.5 is faster-and-cheaper still for focused tasks.

4. Do I need to analyze more than 200k tokens of context?

   Yes: Use Opus 4.7, Sonnet 4.6, GPT-5.5, or Gemini 3 Pro — all support 1M context.

5. Am I working with images, screenshots, or diagrams?

   Yes: Opus 4.7 (high-resolution 2576px support), Gemini 3 Pro (native multimodal), or GPT-5.5 (computer use).

6. Am I stuck and the current model keeps making the same mistake?

   Yes: Switch model family. /best-of-n is the fastest way to try three options at once.

Switching Models in Cursor

The model picker is in the agent panel, right next to the mode selector. You can switch models mid-conversation — the new model picks up the existing context. In the Cursor 3.0 Agents Window you can run different models per agent tab simultaneously.

Keyboard shortcut: Press Cmd+. (macOS) or Ctrl+. (Windows/Linux) to quickly cycle through modes. For model selection, click the model name in the agent panel.

Auto Mode (Composer 2 Default)

Auto mode in Cursor 3.x defaults to Composer 2. When a request is complex enough that Composer 2 would struggle, Auto routes to Opus 4.7 or Sonnet 4.6 automatically. Auto uses blended pricing ($1.25 input, $6.00 output per 1M tokens) and handles model degradation — if one provider is slow, it reroutes.

For beginners, Auto is a reasonable starting point. Once you develop a feel for which model suits which task, manual selection gives you more control and often better results.

Cost Optimization in Practice

Typical Monthly Usage Patterns

Developer Style	Primary Model	Monthly Cost Estimate
All Opus 4.7 / xhigh	Claude Opus 4.7	$100-200
Mixed (recommended)	Composer 2 (Auto) for routine, Opus 4.7 for complex	$40-100
Budget-conscious	Composer 2 + Sonnet 4.6 + Haiku 4.5	$25-60

Cost-Saving Strategies

1. Let Auto pick Composer 2 by default — it’s the cheapest fast-quality option for iteration
2. Start conversations fresh — long conversations accumulate context that costs money on every message
3. Use @ references instead of pasting large blocks of code — Cursor handles file references more efficiently
4. Reserve Opus 4.7 xhigh for genuinely hard problems — the higher effort level spends materially more tokens
5. Enable Max mode only when needed — do not leave it on permanently

When This Breaks

Model seems to have gotten worse: Models do not regress, but API performance varies. If a model is producing lower-quality output than usual, try the same prompt 10 minutes later or switch to a different model temporarily.

Switching models mid-conversation loses context: Rare but can happen with very long conversations. If you notice degraded quality after a switch, start a new chat with the new model and @-reference specific files.

Auto mode keeps selecting a model you dislike: Disable Auto and select manually. The two seconds it takes to choose a model are worth the consistency.

Usage runs out before month end: Check your usage in Settings. If you are burning through Opus 4.7 tokens on tasks that Sonnet 4.6 or Composer 2 could handle, shift your default for routine work.

Opus 4.7 responses feel flat: Opus 4.7 follows instructions more literally and has a shorter default response. If you need the warmer verbose style of Opus 4.6, you can pin Opus 4.6 as a legacy option — but most users adapt quickly by raising the effort level to xhigh.

What’s Next

Project Rules Set up .cursor/rules/ so the AI generates code that matches your team's conventions from the very first prompt.