Claude Fable 5 faces growing backlash after its July 1 re-release, with users claiming stricter guardrails have crippled the flagship model's coding, debugging, and agentic performance. According to BridgeMind, the benchmark group re-ran the July 1 version and recorded sharp declines across key metrics. Debugging fell from 86.2 to 25.9, refactoring dropped from 73.6 to 38.4, and hallucination handling slipped from 75.9 to 61.7. The mechanics behind these numbers reveal that only three of 12 debugging tasks were completed without falling back to Claude Opus 4.8, with every fallback scoring zero. BridgeMind stressed that the model did not get worse - it got caged, indicating the performance decline reflects blocked tasks rather than weaker reasoning capabilities.
Anthropic has restored global access to Claude Fable 5 on Thursday, July 2, following the US government's decision to lift export controls that had forced the company to suspend the model worldwide since June 12. As reported by Anthropic, the company confirmed the restoration across multiple platforms, stating "Fable 5 became available again from Thursday across the Claude Platform, Claude.ai, Claude Code and Claude Cowork, with the company saying it would restore access on AWS, Google Cloud and Microsoft Foundry 'as quickly as possible.'" The restoration comes after Anthropic suspended access for foreign nationals worldwide across every product surface when the export controls were initially imposed, taking both models entirely offline for customers on Claude.ai, the API, AWS Bedrock, and other partner platforms until the situation could be resolved.
The US Commerce Department has lifted export controls on Anthropic's Claude Fable 5 and Mythos 5 models effective July 1, ending nearly three weeks of tense negotiations between the company and the Trump administration. As reported by Anthropic, the company confirmed the reversal directly on June 30, stating "We've received notice that the Department of Commerce has lifted export controls on Claude Fable 5 and Mythos 5. We'll begin restoring access tomorrow, and will share an update soon." The original directive had been imposed on June 12, just three days after the models launched on June 9, citing national security concerns tied to potential model jailbreaks. According to CNN, a trusted partner identified as Amazon had found a way to circumvent Fable's safeguards, though Anthropic had disputed the severity of the flaw at the time.
The lifting of export controls has spurred closer collaboration between Anthropic and US agencies, with the company deepening its partnership through giving designated agencies early access to frontier models and their safeguards before public release, sharing threat intelligence, and working toward a common security standard across AI developers. As reported by CoinDesk, Anthropic is also drafting a framework with Amazon, Microsoft and Google for scoring how dangerous a given jailbreak is, arrangements that tie into a June 2 executive order on AI security and point to a tighter link between frontier AI releases and government review. Alongside the model redeployment, Anthropic said it is increasing cooperation with the U.S. government on model testing, safeguards, and misuse tracking, including pre-release access to models and safety systems for evaluation, information sharing on jailbreaks and misuse, and dedicated resources for joint research. The company has also started drafting a framework with Amazon, Microsoft, Google, and other partners through Project Glasswing, a cybersecurity collaboration announced in April, to assess the severity of AI jailbreaks as a consensus effort for classifying jailbreak risks.
Anthropic's own internal testing found Claude Fable 5 posed no more cybersecurity risk than its competitors, contradicting the original justification for the export controls. As reported by Anthropic, the company tested rival models including Claude Opus 4.8, GPT-5.5, and Kimi K2.7 to gauge the real threat behind the restriction. The findings revealed that every model tested could identify the same vulnerabilities Fable 5 flagged in the Amazon report and reproduce the single exploit demonstration. The admission accompanies Fable 5's global relaunch, capping an 18-day suspension triggered by US export controls on June 12. The finding suggests the directive targeted a gap shared across the industry, not a Fable-specific threat, with Anthropic still building a stronger classifier to block the technique which now flags the reported bypass in over 99% of cases. To resolve the issue, Anthropic worked with the US government to train an 'improved safety classifier' that targets and blocks the specific bypass technique in over 99% of cases, as reported by Anthropic in a post on X. The company had previously noted that the bypass did not expose any unique capabilities specific to Mythos 5.
Restored access carries significant limitations, with Fable 5 drawing from just 50% of weekly usage caps through July 7, then shifting to paid usage credits. The timeline explains the tension, as Anthropic launched Fable 5 on June 9, Washington pulled it offline three days later, and regulators lifted export controls four days after restoring Mythos 5 access for roughly 100 US institutions. The maker of the Claude chatbot was valued at $965 billion after its latest funding round, making it one of the world's most valuable private companies. Anthropic announced June 1 that it had filed confidentially for an initial public offering and is considering going to the public markets as soon as October. However, the shutdown raised questions about the timing of any public listing, with traders wagering on the company's value through a pre-IPO perpetual contract on the onchain exchange Hyperliquid, which fell about 3.7% when access was suspended in June. The company released Fable 5 as the first public-facing version of its Mythos-class model, but with guardrails aimed at containing the full range of its cyber capabilities, preventing it from responding to certain types of queries including those related to cybersecurity and biology, with Claude chatbot routing responses through a different model called Opus 4.8.
