# Claude Fable 5 Is Now Available (and Why We Waited)

We've added Claude Fable 5, Anthropic's most capable model, to AI Data Stream as a premium option. Here's why we left it out until now — and when it's worth using.

**Claude Fable 5** — Anthropic's most capable model — is now available in AI Data Stream. It sits in the model picker as a **premium option**, not the default. This is a deliberate change of heart, so it's worth explaining the decision.

## Why We Left It Out Before

When Fable 5 first became available, we chose **not** to add it. For analytics work — asking questions of your GA4, Search Console, and Google Ads data — Sonnet and Opus already cover the range extremely well. Fable 5's premium pricing didn't buy enough extra value for that kind of task to justify putting it in front of everyone. We'd rather keep the picker focused than pad it with an expensive option most people don't need.

## What Changed

We recently added [OpenAI's full GPT-5.6 lineup](/blog/openai-gpt-5-6-now-available/), including **Sol**, their most capable model. OpenAI positions Sol as close to Fable 5 in raw capability — so with a genuine top-tier OpenAI option now in the picker, it made sense to offer the matching top-tier Claude option alongside it.

Some users specifically want the single most capable model available for the hardest questions, and they should be able to pick it on either side. So Fable 5 is now there for the people who want it.

## Pricing — This One Is Premium

On the standard API we use, Fable 5 is priced well above the rest of the Claude lineup:

- **Input:** $10 per million tokens
- **Output:** $50 per million tokens

That's roughly **2× Opus 4.8** and **3× Sonnet 5**. This is exactly why it's not the default — it's there for when you specifically want maximum capability and accept the cost. Because AI Data Stream is bring-your-own-key, you pay your Anthropic account directly, so you stay in full control of the spend.

## One Thing to Check First

Fable 5 requires **30-day data retention** on Anthropic's side. If your Anthropic organization is configured for zero data retention, Fable 5 requests will be rejected — stick with Opus or Sonnet in that case. Everything else in the picker works the same as always.

## It Is Not the Default

Nothing changes unless you choose it. **Sonnet 5 remains the default** for new providers and is still the best all-round choice for day-to-day analytics. Fable 5 simply appears in the picker, labelled as premium, ready when you want it.

## Effort Levels

Fable 5 supports the full range of effort levels — **low, medium, high, xhigh, and max** — the same as Opus 4.8. Higher effort trades speed and tokens for deeper reasoning, so reach for it on genuinely hard questions and dial it back for routine ones.

## The Full Claude Lineup

- **Claude Fable 5** — Most powerful, premium; for the hardest reasoning
- **Claude Opus 4.8** — Most intelligent, for demanding tasks
- **Claude Opus 4.7** — Previous-generation Opus
- **Claude Sonnet 5** — Best balance of speed and capability (default)
- **Claude Sonnet 4.6** — Previous generation
- **Claude Haiku 4.5** — Fastest, great for quick queries

## Switching Models

1. Go to **Settings > AI Providers**
2. Click your Anthropic provider
3. Select Fable 5 from the dropdown
4. Save

Or select it from the model picker when starting a new conversation.

## Questions?

If you have any questions, [contact us](/contact/) or check our [AI providers documentation](/docs/ai-providers/).