Claude Sonnet vs Opus: Which Should You Use?
Beginner guide to Claude Sonnet 5.5 vs Opus 5.5: side-by-side API prices, per-prompt cost math, clear rules on when to upgrade, and 4-step DSH setup.
On this page

Every beginner using Anthropic's Claude models runs into the exact same question: "What is the real difference between half-price Claude Sonnet and double-price Claude Opus—and if Sonnet misses one bug, should I switch everything to Opus forever?"
Short answer: keep Sonnet for 90% of your daily work, and save Opus for multi-file headaches. The table and 4-step setup below show you exactly when to switch so you neither overpay nor waste an afternoon retrying broken code.
Side-by-side comparison: Sonnet 5.5 vs Opus 5.5 (plus Fable & GPT)
Using Anthropic's official model overview and official Claude API pricing, here is how the lineup compares (rates per 1 million tokens, roughly 750,000 English words):
| Model Name | Input / Output Price | Typical 1-Turn Coding Cost | Reply Speed | Plain-English Role |
|---|---|---|---|---|
| Claude Sonnet 5.5 | $2.00 / $10.00 | ~$0.04 (4 cents) | Fast, snappy | Everyday default: single-file scripts, UI styling, summaries, routine bugs |
| Claude Opus 5.5 | $4.00 / $20.00 | ~$0.08 (8 cents) | Moderate | Multi-file workhorse: changes touching 4+ modules, subtle race conditions |
| Claude Fable 5.2 | $8.00 / $40.00 | ~$0.16 (16 cents) | Slower (deep thinking) | Top-tier specialist when even Opus gets stuck on long open-ended engineering |
| GPT-6.2 Sol (comparison) | $2.00 / $10.00 | ~$0.04 | Fast | OpenAI's everyday coder in the exact same $2 / $10 price bracket |
(Note: the typical 1-turn figure assumes 10,000 input tokens + 2,000 output tokens; older blog posts like TokenMix's April 2026 Sonnet 4.6 vs Opus 4.7 guide reflect older price gaps, so use the 5.5 numbers above today.)
Plain-English rule: rule out setup mistakes before blaming Sonnet
Half the time beginners think "Sonnet isn't smart enough," the real culprit is missing context that would trip up Opus too:
- Check that the AI can actually see your files: Make sure you clicked
Choose workspaceon the right project folder and pasted the full error message before paying 2x for a bigger model. - Keep quick, easy-to-check jobs on Sonnet 5.5: Converting data formats, tweaking CSS layouts, writing unit tests for one function, or summarizing docs all run great at half price.
- Switch to Opus 5.5 when changes span 4+ connected files, touch sensitive auth/concurrency logic, or fail twice on Sonnet: As explained in Anthropic's cost-and-intelligence guide, four failed Sonnet retries burn the same money as one clean Opus fix (see our coding agent task-cost breakdown).
4-step beginner setup: enable both Sonnet and Opus in DSH
You do not have to commit to just one model. Enable both in five minutes and switch whenever a task demands it:
Step 1: Separate web subscriptions from API keys, and grab an sk-ant-... key
Consumer claude.ai subscriptions and developer API keys use separate wallets (as noted in Claude Code's cost docs and Claude Code model config docs, estimated session dollar counters shown to subscription users are not API bills).
To connect DSH, sign in at console.anthropic.com/settings/keys, add at least $5 under Billing, and create a new sk-ant-... secret key.
Step 2: Choose the built-in "Anthropic" card in DSH
Open DSH on your computer and click Settings -> Models -> Add provider:
- Select the built-in
Anthropiccard, paste yoursk-ant-...key, and click Save (full walkthrough in our DSH Anthropic Claude guide). - Avoid this trap: Never paste an
sk-ant-...key into an OpenAI-compatible custom form—Claude requires theanthropic-messagesprotocol handled automatically by the built-in Anthropic card.
Step 3: Turn on both Sonnet 5.5 and Opus 5.5
Enable both Claude Sonnet 5.5 and Claude Opus 5.5 in your Anthropic model list, and pick Claude Sonnet 5.5 as your everyday starting model.
Step 4: Escalate in a fresh session when Sonnet gets stuck
Do your normal coding on Sonnet 5.5. Whenever Sonnet misses a fix twice, click New session, switch the model picker to Claude Opus 5.5, and paste the clean problem statement there (starting a fresh chat avoids dragging failed guesses into the new turn; to test both models systematically on your own repo, use our coding benchmark template).
Common questions
Does my $20/month Claude Pro subscription pay for API keys in DSH?
No. claude.ai Pro/Max subscriptions and console.anthropic.com API credits are billed separately. DSH's Anthropic API-key card draws from your prepaid developer API balance.
Does Opus 5.5 always cost twice as much per finished bug fix?
Not when it saves retries. Four failed turns on $2 / $10 Sonnet burn the same money—and much more of your time—than one clean pass on $4 / $20 Opus.
What if even Opus 5.5 gets stuck on a massive refactor?
For open-ended, multi-repository tasks where Opus still struggles, Anthropic offers a higher $8 / $40 tier called Claude Fable 5.2 (compared side by side in Opus 5.5 vs Fable 5.2).
Model and workflow guides
Gemini 4 Neon is here. When is it worth switching models?
Compare Gemini 4 Neon in two coding setups, work through a $0.45 API example, and find Neon in Gemini, Google AI Studio, and Antigravity.
Read →Models and costsGPT-6.2 Astra & Sol Guide: Specs, Pricing & Choice
Beginner guide to OpenAI's GPT-6.2 Astra and Sol: Claude vs DeepSeek price table, plain-English benchmarks, 272K cliff, and step-by-step DSH & Python setup.
Read →Models and costsGPT-6.1 Terra Guide: Uses, Pricing and Sol Comparison
Beginner guide to OpenAI's GPT-6.1 Terra: half-price visual, computer-use, and writing model compared with Sol and DeepSeek, plus DSH and Python setup steps.
Read →Model setup and costsConnect Local Qwen 3.8 27B to DSH
Beginner guide to connecting local Qwen 3.8 27B to DSH: GPU memory table, LM Studio / Ollama / llama.cpp port setup, exact 5-field config, and error fixes.
Read →Model setup and costsCodex Costs: API Billing vs Subscription Usage
Beginner guide to OpenAI Codex costs: ChatGPT Plus login vs API-key billing table, real per-prompt dollar math, and 3 steps to lock a monthly budget cap.
Read →Model setup and costsConnect Grok API to DSH: Setup & Tools
Step-by-step beginner guide to connecting xAI's Grok API (grok-4.7) to DSH: Grok vs Groq comparison, exact 5-field custom provider setup, and 401/404 fixes.
Read →Model setup and costsChoose a Gemini Model by Task and Cost
Beginner guide to Google Gemini model selection: compare 3.8 Flash, 4 Argon, and DeepSeek prices, watch 2027 rate changes, and set up DSH in 4 simple steps.
Read →Model benchmarks and engineeringGemini 4 Argon Coding Benchmarks & 1M Window
Beginner guide to Google's Gemini 4 Argon coding scores vs Claude and GPT, how its shared 1M window and 65K output cap work, plus 4-step DSH setup.
Read →API costs and enterprise ROIGemini 4 Argon API Pricing, Cache & ROI Guide
Beginner guide to Gemini 4 Argon API pricing: Dec 2026 2x rate jump, how 95% auto-caching and thinking tokens bill, DeepSeek cost comparison, and calculator.
Read →Models and costsOpus 5.5 vs Fable 5.2: Which Model Is Right for You?
Beginner guide to Claude Opus 5.5 vs Fable 5.2: side-by-side API prices, coding benchmarks, per-task cost math, 30-day retention rules, and 4-step DSH setup.
Read →Content creationHow to draft a video script with Media Desk for DeepSeek Harness
Turn a phone note into a video script, cover, and storyboard with Media Desk for DeepSeek Harness. See how to install it, set up images, and export a draft.
Read →Mobile and remote accessUse DeepSeek Harness on Your Phone: Beginner Guide
Zero-jargon beginner guide to using DeepSeek Harness on your phone: find your computer IP, connect over home Wi-Fi in 2 minutes, or use Tailscale on 4G/5G.
Read →Model interfaces and workflowsMiniMax H3 vs M-Series in DSH: Which Works?
Beginner guide to MiniMax H3 vs M3 in DSH: why H3 is for video while M3 handles coding, how to avoid region 401 errors, and 4-step DSH setup instructions.
Read →API credentials and troubleshootingKimi API Key Errors: Region, Model and Permissions
Beginner fix guide for Kimi / Moonshot API key errors: match .ai vs .cn URLs, avoid retired model names and K3 Max typos, decode 401/404/429, and set up DSH.
Read →Image editing and toolsAI Image Editing: Prompts, Masks and Review
Beginner guide to AI image editing without ruining your subject: copy our 4-line prompt template, avoid PNG mask traps, inspect edges, and use DSH safely.
Read →