Claude Opus vs Sonnet vs Haiku comes down to the job: Anthropic’s own model guide points different kinds of work to different models. Tell the picker below what you’re doing and how you use Claude, and you get the model to start with, the reasons, whether your plan includes it, and what the job costs on each model through the API.
It covers all four current models, Claude Fable 5.1, Opus 5.5, Sonnet 5.5 and Haiku 5.5, and runs entirely in your browser.
Free model picker
Opus, Sonnet or Haiku: which Claude model fits the job?
Tell us the job, how you use Claude and what matters most. You get the model to start with and why, whether your plan includes it, and what the job costs on each model through the API. Nothing you enter leaves your browser.
Where do you use Claude?
What's the job?
What matters most?
Your Claude plan
Skip this if you only use the API.Get an alert when Claude prices change
Join the weekly HOW TO // AI newsletter, where we flag AI price and limit changes, and unlock the advanced cost scorecard: your job's cost on every Claude model with prompt caching, batch, US-only inference and fast mode.
Unlocked: the advanced cost scorecard is below. You are also on the weekly HOW TO // AI newsletter.
Free. You also get the HOW TO // AI newsletter once a week. Unsubscribe any time.
What your job costs on each model (API)
Using the Claude app on a plan? You pay the plan price, not per token, so skip this part. It's for API users.
How we calculate: cost per request = input tokens / 1,000,000 x input price + output tokens / 1,000,000 x output price. Monthly cost = cost per request x requests a day x 30. Haiku 5.5 charges higher prices when a prompt is over 100,000 tokens, and we apply those automatically. Standard global prices, before prompt caching. Claude 5.x models all use Anthropic's newer tokenizer, so the same text counts the same on each of them.
The current Claude models, side by side
Who should pick which
Sources: Anthropic: Models overview, Anthropic: Choosing a model, Anthropic: API pricing, Claude plans and pricing, Claude Fable models on your plan, checked October 18, 2026. Plan prices from our AI Plan Watch data, checked November 4, 2026. API prices in US dollars per million tokens. The pick follows Anthropic's own model guide; test on your real prompts before you commit.
How to use the model picker
- Pick where you use Claude: the app, the API or both.
- Pick the job and what matters most: speed and cost, a balance, or the best possible answer.
- App users: pick your plan to see whether it includes the model.
- API users: enter requests a day and tokens per request to see the monthly cost on every model.
Opus vs Sonnet vs Haiku at a glance
| Model | Anthropic says it’s for | Speed | API price per 1M tokens (input / output) |
|---|---|---|---|
| Claude Fable 5.1 | Demanding reasoning and long-horizon agentic work | Slower | $10 / $50 |
| Claude Opus 5.5 | Long-running agentic coding and knowledge work | Moderate | $4 / $20 |
| Claude Sonnet 5.5 | The best combination of speed and intelligence | Fast | $2 / $10 |
| Claude Haiku 5.5 | High-volume, latency-sensitive tasks such as classification, extraction and routing | Fastest | $0.10 / $0.50 for prompts up to 100,000 tokens |
All four list a 1M-token context window and up to 128K tokens of output. Haiku 5.5 is the only one priced by prompt length: prompts over 100,000 tokens pay $0.50 per million input tokens and $2.50 per million output tokens. The Batch API halves every price for work that can wait.
When to pick Opus, Sonnet or Haiku
- Sonnet 5.5 for everyday coding, data analysis, writing and content. Anthropic’s guide recommends it for speed and capability on everyday coding, agent and enterprise work.
- Opus 5.5 for long agentic coding jobs, large refactors, complex systems work and computer use. If you’re unsure, Anthropic says to start with Opus 5.5 for most workloads.
- Haiku 5.5 for real-time features and high-volume sorting, extraction or routing, where speed and price matter most.
- Fable 5.1 for agent sessions that run for hours and research that has to end as a finished document, spreadsheet or deck, or when Opus at higher effort still falls short.
Before you switch models, try the effort setting. Anthropic notes that tuning effort is often a better lever than switching models. Opus 5.5 and Haiku 5.5 default to medium effort; Sonnet 5.5 and Fable 5.1 default to high.
Which Claude plans include which model
| Plan | Haiku | Sonnet | Opus | Fable |
|---|---|---|---|---|
| Free | Yes | Yes | No | No |
| Pro | Yes | Yes | Yes | Pay-as-you-go usage credits |
| Max 5x and 20x | Yes | Yes | Yes | Up to 50% of weekly limits |
| Team | Yes | Yes | Yes | Premium seats up to 50% of weekly limits; standard seats use usage credits |
Plan prices change, so the picker reads them live from our AI Plan Watch data. For the full plan breakdown, see Claude pricing and Claude Pro vs Max.
Common mistakes
- Paying Opus prices for Sonnet work. On the API, Opus 5.5 costs twice what Sonnet 5.5 does per token. Test the cheaper model on your real prompts first.
- Sending huge prompts to Haiku. Above 100,000 tokens, Haiku 5.5’s input price goes up five times.
- Forgetting batch and caching. Results that can wait are half price with the Batch API, and repeated prompt text is billed at a fraction of the input price as a cache hit.
- Picking a model your plan doesn’t have. The free Claude plan includes Sonnet and Haiku only. Opus starts on Pro.
Related on HOW TO // AI
- Claude pricing: every Claude plan, with its limits.
- Claude Pro vs Max: which subscription is worth it for you.
- How to use Claude Code: a step-by-step beginner’s guide.
- Claude certification: the four Anthropic exams and who can take them.
Checked against Anthropic’s models overview, choosing-a-model guide and API pricing page, and claude.com/pricing, on October 18, 2026.
Frequently asked questions
Is Claude Opus better than Sonnet?
Opus 5.5 is Anthropic's pick for long-running agentic coding and complex work, while Sonnet 5.5 is pitched as the best combination of speed and intelligence. Anthropic's guide points everyday coding, data analysis and content creation to Sonnet, which costs half as much as Opus per token on the API.
Which Claude model is the cheapest?
Claude Haiku 5.5. It costs $0.10 per million input tokens and $0.50 per million output tokens for prompts up to 100,000 tokens, and $0.50 and $2.50 above that (Anthropic pricing, October 18, 2026).
Can I use Opus on the free Claude plan?
No. The free plan includes Sonnet and Haiku. Opus is included from Pro up, and Fable is available on Pro through usage credits and on Max for up to 50% of your weekly limits.
What is Claude Fable 5.1?
Anthropic's most capable model open to all customers, built for demanding reasoning and long-horizon agentic work. On the API it costs $10 per million input tokens and $50 per million output tokens.
Do all Claude models have the same context window?
The four current models all list a 1M-token context window and up to 128K tokens of output. What changes is speed, price and how much reasoning they bring to hard tasks.