Claude Fable 5: Anthropic's Most Powerful AI Model -- Free to Try Until June 22

On June 9, 2026, Anthropic publicly released its most powerful AI model to date: Claude Fable 5. It is the first generally available Mythos-class model and outperforms the previous flagship Opus 4.8 on nearly all benchmarks -- by over 10 percentage points in some cases. The best part: Pro subscribers can try Fable 5 for free until June 22.

What is Claude Fable 5?

Fable 5 is a safeguarded version of Anthropic's Mythos 5 model. Mythos 5 itself is only available through the limited-access research program Project Glasswing, as it is too capable in certain areas like cybersecurity and biochemistry. Fable 5 retains the core strengths -- deep reasoning, long autonomous work, excellent code comprehension -- but automatically falls back to Opus 4.8 for high-risk queries. The result: frontier model performance with built-in safety guardrails.

Why Fable 5 is a real upgrade:

1. Longer autonomous work sessions

Fable 5 can work independently on a task for significantly longer than previous models. Instead of losing the thread after a few steps, it plans multi-step tasks, asks clarifying questions, and works through them step by step. This makes it ideal for tasks where you previously had to constantly intervene.

2. Superior code comprehension

The numbers are impressive: 80.3% on SWE-bench Pro (vs. 69.2% for Opus 4.8) -- a benchmark that measures how well a model finds and fixes real software bugs in large codebases. Stripe reports that Fable 5 completed a codebase-wide migration in one day that would have taken an entire team over two months.

3. Better document understanding

Fable 5 understands diagrams, tables, and charts in PDFs and documents better than any previous Claude model. This opens new possibilities for financial analysis, legal research, and scientific work.

4. 1 million token context window

Fable 5 can process up to 1 million tokens simultaneously -- roughly 750 pages of text. And it can output up to 128,000 tokens at once, which is especially helpful for lengthy analyses or code generation.

What to use Fable 5 for:

- Complex research: 'Analyze these 15 studies on topic X. Find contradictions, identify research gaps, and create a structured summary with key findings.'
- Code projects: Large refactoring, migrations, and architecture reviews where the model needs the full project context
- Document analysis: Comparing contracts, evaluating financial reports, working through scientific papers -- including embedded tables and graphics
- Long-running tasks: Multi-hour agent-based sessions where Fable 5 independently plans, executes, and verifies results

When Opus 4.8 remains the better choice:

Not every task needs Fable 5. For quick questions, short texts, brainstorming, and everyday chat conversations, Opus 4.8 is the smarter choice -- it is faster, cheaper, and equally capable on simple tasks. The rule of thumb: the more complex and lengthy the task, the greater Fable 5's advantage.

Availability and pricing:

- Free trial: Pro, Max, Team, and Enterprise users can use Fable 5 at no extra cost from June 9 to 22, 2026. It counts as 2x usage (one Fable 5 message consumes as much as two Opus messages)
- After June 23: Fable 5 will be billed via Usage Credits -- a balance you top up in addition to your subscription
- API pricing: 10 USD per million input tokens and 50 USD per million output tokens (2x Opus 4.8). Prompt caching with 90% discount on cached input tokens
- Platforms: Claude.ai, Claude Code (terminal), Amazon Bedrock, Microsoft Foundry, GitHub Copilot
- Model ID: claude-fable-5

How to activate Fable 5 now:

1. Open claude.ai and sign in
2. Click the model picker at the top of the chat (it likely says 'Claude Opus 4.8')
3. Select 'Claude Fable 5' from the list
4. Test it with a demanding task -- you will barely notice a difference on simple questions

In Claude Code (terminal): Type /model and select Fable 5 from the menu.

Fable 5 vs. Opus 4.8 comparison:

| | Opus 4.8 | Fable 5 |
|---|---|---|
| Strength | Fast and reliable | Deep reasoning, autonomous work |
| Context window | 200,000 tokens | 1,000,000 tokens |
| Max output | 16,000 tokens | 128,000 tokens |
| SWE-bench Pro | 69.2% | 80.3% |
| Ideal for | Daily use, chat, short tasks | Complex projects, research, code |
| API cost | 5 / 25 USD per 1M tokens | 10 / 50 USD per 1M tokens |
| Subscription usage | 1x | 2x |

An important note about safety guardrails:

Fable 5 includes safety classifiers that automatically fall back to Opus 4.8 for certain high-risk queries (cybersecurity exploits, biological/chemical synthesis, medical diagnostics). This means you do not get full Fable 5 performance in these areas, but you get a more reliable and safer model overall. For everyday work -- writing, code, analysis, research -- the guardrails are not noticeable.

Privacy note:

Anthropic's privacy policies apply to Fable 5 just as they do to all Claude models: Pro and Team conversations are not used for model training. For enterprises, Anthropic offers the option to process data in European data centers through Amazon Bedrock and Microsoft Foundry. If you are processing sensitive company data, use the API or Enterprise version with appropriate data processing agreements.

What this means for you:

Claude Fable 5 is not an incremental update -- it is a generational leap. Tasks where previous models would give up after a few steps or lose context can now be worked through independently and persistently by Fable 5. The free trial period until June 22 is the perfect opportunity to test the model with a demanding task: upload a complex document, give a multi-step assignment, or let it work on a real coding problem. That way you will quickly see whether the upgrade is worth it for your workflow.

Sources: anthropic.com/news/claude-fable-5-mythos-5, techcrunch.com/2026/06/09/anthropic-released-claude-fable-5-its-most-powerful-model-publicly-days-after-warning-ai-is-getting-too-dangerous, truefoundry.com/blog/claude-fable-5-vs-opus-4-8-benchmarks-pricing-when-to-use-each, ai.rs/ai-for-business/claude-fable-5-free-test-window-how-to-switch