Claude Fable 5 Is Back (July 1): World's Most Powerful AI Returns After US Ban (2026)
The US placed export controls on Anthropic's Claude Fable 5 — the most capable AI ever — forcing a global suspension. Today (July 1) the controls are lifted and Fable 5 is back. Here's what it can do.
Anthropic launched Claude Fable 5 on 9 June 2026 — its most capable model ever, state-of-the-art on nearly all tested benchmarks. Just three days later, on June 12, the US government applied export controls to Fable 5 (and its sibling Mythos 5), requiring Anthropic to restrict access to foreign nationals. Because the order took effect immediately and there was no way to verify nationality in real time, Anthropic suspended access for everyone. On June 30, the export controls were lifted — and per Anthropic's official 'Redeploying Fable 5' notice, the model is available again globally from today, Wednesday July 1. Here's the full story of the most powerful — and most fought-over — AI of 2026.
- Claude Fable 5 is Anthropic's most capable AI — state-of-the-art on nearly all benchmarks.
- June 12: the US placed export controls on it → Anthropic suspended access for everyone.
- The suspension blocked all users, including Indian developers and creators.
- June 30 the controls were lifted — Fable 5 is back globally from today, July 1 2026.
- It has a 1M-token context, always-on 'adaptive thinking', vision and full tool use.
The story: a two-week ban, then back online
When Anthropic launched Fable 5 on 9 June, it warned the model is so capable that, without safeguards, its abilities "in areas like cybersecurity could be misused to cause serious damage" — so it shipped with safety classifiers that refuse dangerous queries. Just three days later, on 12 June, the US government applied export controls to Fable 5 and its sibling Mythos 5, requiring access to be restricted to foreign nationals. Because the order was immediate and Anthropic had no way to verify nationality in real time, it suspended access for every user — cutting off Indians and other foreigners and sparking a wider debate about AI sovereignty. The suspension held for about two-and-a-half weeks. Then, on 30 June, the export controls were lifted, and Anthropic's official Redeploying Fable 5 notice confirmed the model is available again globally from today, 1 July — on the Claude Platform, Claude.ai, Claude Code and Claude Cowork. Bottom line: the most powerful AI of 2026 was briefly too hot to stay online — and now it's back.
- 19 June — Launch
Fable 5 ships as Anthropic's most capable model
- 212 June — US controls
Export controls force a global suspension
- 312–30 June — Ban
Blocked for all users, including India
- 430 June — Lifted
US removes the export controls
- 51 July — Back
Available again globally on Claude
- 19 June — Launch
Fable 5 ships as Anthropic's most capable model
- 212 June — US controls
Export controls force a global suspension
- 312–30 June — Ban
Blocked for all users, including India
- 430 June — Lifted
US removes the export controls
- 51 July — Back
Available again globally on Claude
What is Claude Fable 5?
Claude Fable 5 (claude-fable-5, launched 9 June 2026) is Anthropic's most capable widely-released model, built for the hardest reasoning and long-horizon agentic work — sitting above Claude Opus 4.8. It runs with adaptive thinking always on, a 1M-token context window, vision, and a full agentic toolkit.
- – Hardest science & math
- – Multi-step logic
- – Always-on adaptive thinking
- – Whole-repo refactors
- – Long-horizon autonomous tasks
- – Code execution
- – Programmatic tool calling
- – Memory tool
- – Context editing & compaction
- – 1M tokens (~555k words)
- – Whole books / codebases
- – 128k output
- – Read images, charts, docs
- – Multimodal analysis
- – Built-in refusal classifiers
- – Fallback to other Claude models
How powerful is it? (benchmarks)
These are Anthropic's own published numbers for Fable 5 (reported alongside Mythos 5). On agentic coding — the headline test — Fable 5 leads every rival by a wide margin:
📊 Agentic coding — SWE-Bench Pro (success rate %)
Anthropic's most capable model
📊 FrontierCode (Diamond) — hardest coding tasks, xhigh effort (%)
more than 2× the next best
Full benchmark comparison
📊 Claude Fable 5 vs the frontier
| ★ WinnerClaude Fable 5 | Claude Opus 4.8 | GPT-5.5 | Gemini 3.1 Pro | |
|---|---|---|---|---|
| Agentic codingSWE-Bench Pro | 80.3% | 69.2% | 58.6% | 54.2% |
| Agentic codingFrontierCode (Diamond) | 29.3%xhigh | 13.4%xhigh | 5.7%xhigh | — |
| Agentic codingTerminal-Bench 2.1 | 88.0% | 82.7% | 83.4%Codex CLI | 70.7%Gemini CLI |
| Knowledge workGDPval-AA | 1932 | 1890 | 1769 | 1314 |
| Knowledge work (vision)GDP.pdf, no tools | 29.8% | 22.5% | 24.9% | 16.7% |
| Spatial reasoningBlueprint-Bench 2 | 38.6% | 14.5% | 36.2% | 26.5% |
| Tool useAutomationBench | 17.4% | 15.5% | 12.9% | 9.6% |
| Computer useOSWorld-Verified | 85.0% | 83.4% | 78.7% | 76.2% |
| Multidisciplinary reasoningHumanity's Last Exam, with tools | 64.5%* | 57.9% | 52.2% | 51.4% |
| CybersecurityExploitBench (Cap %) | 78.0%* | 40.0% | 34.0% | — |
| HealthHealthBench Professional | 66.0%* | 56.9% | 51.8% | — |
| LegalLegal Agent Benchmark | 13.3% | 10.4% | 2.1% | 0.0% |
Fable 5 vs the other frontier models (price & context)
| Model | Context | API price /1M (in/out) | Best at |
|---|---|---|---|
| Claude Fable 5 | 1M | $10 / $50 | Most capable — hardest reasoning + agentic coding |
| Claude Opus 4.8 | 1M | $5 / $25 | Top coding, careful writing |
| GPT-5.5 | ~1M | $5 / $30 | Versatile all-round flagship |
| Gemini 3.1 Pro | ~1M | $2 / $12 | Best value + India INR billing |
What can Fable 5 actually do?
- Run multi-hour autonomous coding tasks — plan, write, test and fix across a whole repo.
- Reason through the hardest science, math and analysis with always-on thinking.
- Use tools programmatically, remember across steps, and execute code.
- Read and reason over images, charts and long documents (1M-token context).
- Decline unsafe requests via built-in classifiers, with automatic fallback to another Claude.
How to use Claude Fable 5 (from India)
- 1Open Claude
claude.ai or the Claude API
- 2Pick Fable 5
Select the model / claude-fable-5
- 3Give your task
A hard question or a multi-step job
- 4It thinks & acts
Adaptive thinking + tools, autonomously
- 5Get the result
Answer, code or completed workflow
- 1Open Claude
claude.ai or the Claude API
- 2Pick Fable 5
Select the model / claude-fable-5
- 3Give your task
A hard question or a multi-step job
- 4It thinks & acts
Adaptive thinking + tools, autonomously
- 5Get the result
Answer, code or completed workflow
Pros
- The single most capable widely-available AI right now — top reasoning + agentic coding.
- Huge 1M context, always-on thinking, memory and code-execution tools.
- Back worldwide (incl. India) after the US lifted its ban.
Cons
- Expensive ($10/$50 per 1M) with no India rupee plan — overkill for everyday tasks.
- Built-in refusal classifiers can decline some requests (plan for fallback).
- Exact public benchmark numbers are still limited at launch.
Save this summary as an image or share it.
AICreatorHub Team
Hands-on AI practitioners covering tools, models and news for India.
%253A%2520Claude%2520Sonne%252C%2520abstract%2520AI%2520tech%252C%2520glowing%2520circuits%252C%2520blue%2520teal%252C%2520no%2520people%252C%2520no%2520text%3Fwidth%3D640%26height%3D360%26nologo%3Dtrue%26seed%3D68721)
