Claude Opus vs Sonnet
Fable, Opus, Sonnet and Haiku: which Claude to use, and when.
Claude's four model families
From deepest to fastest, as Anthropic describes them.
| Family | Built for | Cost to run (newest version, per million tokens in / out) | On claude.ai |
|---|---|---|---|
 Fable | Demanding reasoning and long agent work | $10 / $50 (Fable 5.1) | Pro through usage credits; Max[1],[2] |
| Opus | Long-running coding and knowledge work | $4 / $20 (Opus 5.5) | Pro and up[1],[2] |
| Sonnet | The best mix of speed and intelligence | $2 / $10 (Sonnet 5.5) | Free and up[1],[3] |
| Haiku | The fastest answers | $1 / $5 (Haiku 4.5) | Free and up[1],[2] |
The Claude models in Chatly
With the note Chatly shows for each. Anthropic's September releases are not in Chatly yet.
| Model | Chatly's note |
|---|---|
| Claude Fable 5 | Frontier reasoning & agentic power. |
| Claude 5 Opus | Frontier agentic coding & enterprise work. |
| Claude 5 Sonnet | Frontier coding & agents, Sonnet speed. |
| Claude 4.5 Haiku | Our fastest, most concise model. |
Ask Opus and Sonnet the same thing
Each opens Chatly Chat on that model with the same request ready.
Then put the answer to work
The same plan runs the agent behind the chat. It makes the files, not just the text, and works with Gmail, Slack and 40+ other apps.
Turn an answer into slides
A deck with real charts and speaker notes, ready to present or edit.
Write the document
Reports, proposals, contracts and invoices, formatted and ready to send.
Build the spreadsheet
Budgets, trackers and models that arrive built, filled in and charted.
Publish a website
A landing page or a full site, live on its own link from one request.
Make the images
Product shots, headshots and social posts from the top image models.
Research with sources
Deep research that reads the web for you and cites every claim.
Chatly plans
Three plans, and Teams on top: pick how long you pay for, and how many credits.
Standard
$18/ month, billed quarterly
Perfect for getting started with AI
Subscribe7K credits / month
- Flash and Pro modes
- 3 concurrent agent tasks
- 2 scheduled tasks
- Connect up to 2 apps, like Gmail and Slack
- 2 knowledge bases
- Commercial usage rights
- Build and publish websites
- Create presentation-ready slides
- Deep research with citations
- Generate docs, sheets and audio
- Generate images and videos
Professional
$36/ month, billed quarterly
More credits, more room to create
Subscribe- Flash, Pro and Ultra modes
- 5 concurrent agent tasks
- 5 scheduled tasks
- Connect up to 5 apps, like Gmail and Slack
- 5 knowledge bases
- Commercial usage rights
- Build and publish websites
- Create presentation-ready slides
- Deep research with citations
- Generate docs, sheets and audio
- Generate images and videos
- AI ad creator
- AI video editor
Ultra
$183/ month, billed quarterly
Built for power users who never hit limits
Subscribe- Flash, Pro and Ultra modes
- Unlimited concurrent agent tasks
- 25 scheduled tasks
- Connect all 40+ apps
- 25 knowledge bases
- Commercial usage rights
- Build and publish websites
- Create presentation-ready slides
- Deep research with citations
- Generate docs, sheets and audio
- Generate images and videos
- AI ad creator
- AI video editor
- Early access to new models
Chatly for Teams: pooled credits and shared seats, from three seats. Talk to us
Common questions
What is the difference between Claude Opus and Sonnet?
Anthropic builds Opus for long, complex coding and knowledge work and Sonnet for the best mix of speed and intelligence. Opus costs about twice as much to run as Sonnet through Anthropic's API.
Where does Claude Fable fit?
Fable is Anthropic's model family for the most demanding reasoning and long agent tasks, above Opus. On claude.ai it is reached through paid usage credits on Pro, or the Max plans.
Is Opus better than Sonnet?
For the hardest coding and analysis, usually. For everyday writing and quick work, Sonnet is faster and often just as good. Try both on your own task.
Which Claude models are in Chatly?
Claude Fable 5, Claude 5 Opus, Claude 5 Sonnet, Claude 4.8 Opus, 4.7 Opus, 4.6 Sonnet and 4.5 Haiku. Anthropic's September releases (Fable 5.1, Opus 5.5 and Sonnet 5.5) are not in Chatly yet.
Is Chatly made by Anthropic?
No. Chatly is an independent product with Claude, GPT, Gemini, Grok and other models in one place. It is not affiliated with Anthropic.
Plans, prices and models change, so check each company's own site before you buy.
Ask your own question
The fastest way to choose is to try. Pick one model or two and ask.
Model and product names belong to their makers and are used here only to compare. Chatly is not affiliated with Anthropic, OpenAI, Google, xAI, DeepSeek, Moonshot AI or Perplexity.
Chatly on your phone and your Mac
Chatly for iPhone
Ask Claude, GPT or Gemini on the train, switch models mid-chat, and pick the thread up on your phone. Every chat you start here is on your phone.
Coming soon to the App Store
Chatly for Mac
Every model on your desktop, with the files and screenshots you already have to hand.
Download for macOS