Anthropic just dropped a bomb. Today, June 9, 2026, the company released Claude Fable 5 — the first publicly available model from the Mythos class, the same family that spooked governments worldwide when it demonstrated superhuman ability to find and exploit cybersecurity vulnerabilities.
This is not a routine model update. This is a generational shift.
What Is Claude Fable 5?
Fable 5 is a Mythos-class model — Anthropic's most powerful tier, sitting above the Opus class that has been the ceiling of Claude's capabilities until now. The name comes from the Latin "fabula" (that which is told), a sibling to the Greek "mythos."
In plain terms: Fable 5 is the most capable AI model Anthropic has ever made publicly available. It is state-of-the-art on nearly all tested benchmarks, showing exceptional performance in software engineering, knowledge work, vision, scientific research, and long-horizon autonomous tasks.
The model that preceded it — Claude Mythos Preview — was released in April 2026 but restricted to approximately 50 organizations under Project Glasswing due to its cybersecurity capabilities. Fable 5 is derived from the same training as Mythos but includes additional safety classifiers that block responses in high-risk domains.
The Numbers
The benchmarks are striking. Fable 5 scored 80.3% on SWE-Bench Pro, compared to 69.2% for Opus 4.8 and 58.6% for GPT-5.5. It is the first model to exceed 90% on Hex's analytical benchmark. On extended coding tasks, Fable 5 completed equivalent work with fewer tool calls and lower token consumption than previous Opus-tier models.
| Metric | Fable 5 | Opus 4.8 | GPT-5.5 |
|---|---|---|---|
| SWE-Bench Pro | 80.3% | 69.2% | 58.6% |
| Hex Analytical | 90%+ | ~78% | ~72% |
| Tool efficiency | Best | Good | Average |
Pricing: $10 per million input tokens, $50 per million output tokens — exactly double Opus 4.8. Batch API: $5/$25. The existing 90% input token discount for prompt caching applies, making cached workloads significantly cheaper.
Context window: 1 million input tokens. Output: up to 128K tokens. Model ID: claude-fable-5.
The Safety Architecture
Here's where it gets interesting. Fable 5 includes safety classifiers that actively refuse certain requests. When a query touches high-risk domains — cybersecurity exploitation, biological weapons, chemical synthesis, or model distillation — Fable 5 blocks the response and falls back to Claude Opus 4.8 instead.
The API returns HTTP 200 with stop_reason: "refusal" and a stop_details object naming the triggered category ("cyber", "bio", or "reasoning_extraction"). Crucially, refusals that generate no output are not billed.
A new beta fallbacks parameter lets developers retry a refused request server-side in one round trip, automatically routing to Opus 4.8 when Fable's classifiers trigger. This means your application gracefully degrades rather than failing entirely.
Anthropic says the classifiers trigger in under 5% of sessions — meaning 95%+ of normal usage is unaffected.
Alongside Fable: Claude Mythos 5
Fable 5 isn't alone. Anthropic simultaneously deployed Claude Mythos 5 — the same underlying model but with cybersecurity safeguards removed. Mythos 5 is available only to organizations already approved through Project Glasswing, the program that began in April with partners including Amazon, Microsoft, Apple, Google, and Nvidia.
Mythos 5 is for defensive cybersecurity work. Fable 5 is for everything else.
Availability and Pricing
Fable 5 is available immediately on:
- Claude API (model ID:
claude-fable-5) - Claude Platform & Claude Code
- GitHub Copilot
- AWS Bedrock
- Google Cloud Vertex AI
- Microsoft Foundry
For subscription users, there's a time-limited window: through June 22, Fable 5 is included in Pro, Max, Team, and seat-based Enterprise plans at no extra cost. After June 23, it moves to usage credits only, with plans to restore it as a standard subscription feature as soon as capacity allows.
Why This Matters
Three things make Fable 5 significant beyond the benchmark numbers.
First, it's the public debut of Mythos-class intelligence. Until today, this tier of capability was locked behind government partnerships and enterprise approvals. Now anyone with an API key can access it.
Second, the safety-first architecture is a template for the industry. Rather than choosing between "release everything" and "release nothing," Anthropic found a middle path: release the model with domain-specific circuit breakers that protect the most dangerous capabilities while leaving 95%+ of functionality untouched. The fallback parameter makes this seamless for developers.
Third, the timing. Anthropic is preparing to enter the public markets alongside OpenAI and SpaceX. Fable 5 is both a technical milestone and a statement of competitive position: this is what the most advanced commercially available AI can do.
What This Means for the GCC
For businesses in Saudi Arabia and the Gulf, Fable 5 opens new possibilities. The model's exceptional performance on long-horizon coding tasks makes it ideal for complex software development. Its knowledge work capabilities — sustained reasoning across millions of tokens — are transformative for legal analysis, financial modeling, and strategic planning.
At $10/$50 per million tokens, it's expensive but justifiable for high-value work where quality matters more than cost per token. The prompt caching discount (90% off cached inputs) makes repeated-context workflows — like processing legal documents against a fixed regulatory framework — dramatically more affordable.
Fable 5 is available through AWS Bedrock and Google Cloud Vertex AI, both of which offer Middle East regions for data residency compliance.
This article is published by Conneqt (conneqtme.com) — the AI marketing and automation agency for the GCC and Saudi Arabia.