Anthropic met with the Trump administration on June 15 to discuss U.S. government restrictions on its new Claude Fable 5 and Mythos 5 AI models. Access to both was disabled for all customers on June 12 following an export control directive from the U.S. Commerce Department.
June 12, 2026 · Anthropic
Anthropic Pulls Fable 5 & Mythos 5 Worldwide — Days After Launch
A US Commerce Department export-control directive ordered access suspended for all foreign nationals. Anthropic disabled both models globally, calls it "a misunderstanding," and on June 15 entered talks to restore access.
17:21 ET
Directive received — June 12, citing "national security authorities"
2
Models disabled worldwide; other Claude models (Opus 4.8) unaffected
2×
Roughly double the price of Opus 4.8 ($10 in / $50 out per 1M tokens)
SWE-Bench Pro — software engineering
Higher is better. Fable 5 / Mythos 5 outscore prior frontier models.
FrontierCode Diamond
29.3% vs 13.4% (Opus 4.8)
GDP.pdf (vision)
29.8% vs 24.9% (GPT-5.5)
Why the shutdown
Concerns the models could be misused in cybersecurity and biology — though the government named no specific concerns.
Vulnerability flagged by an Amazon researcher + discovered jailbreak techniques bypassing guardrails
Directive suspends access for all foreign nationals — inside or outside the US, including foreign staff
Ongoing litigation over a Pentagon supply-chain risk designation (blacklist)
Developer praise
"Game changer" for long-horizon agentic tasks
Unlocked previously hard long-term problems (e.g. with Cursor)
Stronger structural understanding in SVG generation
Criticism
Tendency to "over-engineer" simple tasks
Does performance justify ~double the cost?
Opaque regulatory process; may push users toward local models
Continue reading The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.
Already purchased? Sign in ✓ Signed in — this article isn’t included in your current plan.Unlocking the full article…