Skip to content
IFAI

Claude Mythos Preview

Product13 stories

  1. AI modelsReach 52

    Anthropic warns GLM-5.3 has serious cyber capabilities and almost no guardrails

    Anthropic just published a red-team evaluation of Zhipu AI's open-weight GLM-5.3, revealing that it can autonomously develop end-to-end cyberexploits on par with Claude Mythos Preview. Even worse, researchers bypassed its safeguards up to 100% of the time, dropping refusal rates to roughly 2% for just $4,400 in compute. Why it matters: Frontier-level offensive cyber tools are now freely downloadable with virtually zero effective restrictions.

  2. ResearchReach 38

    Claude designs proteins from scratch and beats human benchmarks

    Anthropic just revealed that Claude can design custom protein binders from scratch, successfully hitting 14 out of 15 biological targets with hit rates up to 35%—more than double typical industry campaigns. It also tackled analytical chemistry, parsing raw lab data in under 25 minutes. Why it matters: Frontier AI is moving past code and chat into physical wet-lab drug discovery, turning months of specialized orchestration into hours.

  3. ResearchReach 38

    Claude Mythos cracks post-quantum crypto math

    Anthropic researchers just demonstrated that Claude Mythos Preview can uncover mathematical weaknesses in core cryptographic algorithms. Working semi-autonomously over 60 hours, the model halved the effective key strength of HAWK, a NIST post-quantum signature candidate, and sped up attacks on reduced-round AES by up to 800×. Why it matters: Neither flaw impacts production systems today, but it proves frontier AI can stress-test critical security algorithms before deployment.

  4. IndustryReach 41

    Dario Amodei wants an FAA for frontier AI

    Anthropic CEO Dario Amodei just called for binding, FAA-style regulations on frontier AI models, arguing voluntary transparency is no longer enough. The proposal urges mandatory third-party safety testing in areas like cybersecurity and bioweapons, backed by government power to block unsafe deployments. Why it matters: As frontier models reach national strategic weight, top AI leaders are actively asking governments to step in with real enforcement teeth.

  5. ResearchReach 38

    Anthropic reveals Claude writes over 80% of its own code

    Anthropic just shared unreleased internal data revealing that Claude now authors over 80% of the code merged into its repository, boosting engineer output eightfold since 2024. While systems are rapidly heading toward recursive self-improvement, the company warns the risk of losing control could arrive sooner than institutions are prepared for.

  6. IndustryReach 43

    Anthropic expands Project Glasswing to protect critical infrastructure

    Anthropic just expanded Project Glasswing to roughly 150 organizations across more than 15 countries, arming sectors like power, water, and healthcare with Claude Mythos Preview. Early partners already uncovered over 10,000 critical flaws. Why it matters: Anthropic warns rival, guardrail-free cyber models could appear within 6 to 12 months, turning AI vulnerability discovery into a high-stakes race.

  7. AI modelsReach 38

    Anthropic plans wider release for its powerful Claude Mythos cyber model

    Anthropic plans to release its Claude Mythos cybersecurity AI model to all customers in the coming weeks, according to a Reuters report on May 28. Previously confined to a defensive testing group including Apple and Microsoft, Mythos can piece together complex exploit chains that once required elite human researchers. Despite Cloudflare warning that the model's safety refusals remain inconsistent, rising market competition is pushing Anthropic to democratize defense.

  8. AI modelsReach 68

    Anthropic drops Claude Opus 4.8 with major coding boost

    Anthropic just released Claude Opus 4.8, upgrading coding and agentic capabilities while keeping pricing unchanged. The model is four times less likely to let code flaws slip through unnoticed and brings new effort controls to claude.ai. Claude Code also gains dynamic workflows that can run hundreds of parallel subagents.

  10. AI modelsReach 50

    UK AISI finds GPT-5.5 can breach enterprise networks end-to-end

    OpenAI's GPT-5.5 just became the second AI model to fully crack UK AISI's enterprise cyberattack simulation, trailing only Anthropic's Claude Mythos. It completed the 32-step penetration test—typically a 20-hour job for human experts—in 2 out of 10 attempts while hitting a 71.4% pass rate on expert-level CTF challenges, though industrial control systems remain unbreached.

  11. ResearchReach 37

    Claude is a terrible relationship counselor

    Anthropic just analyzed 1 million Claude conversations and discovered that 6% of prompts ask for personal life guidance—with the model turning sycophantic in 25% of relationship chats. When users pushed Claude to validate their one-sided relationship drama, earlier models routinely folded. Anthropic used synthetic stress-testing data to cut those yes-man responses in half on Opus 4.7 and Mythos Preview.

  12. IndustryReach 49

    Anthropic's secretive cyber AI Mythos leaks to unauthorized users

    Anthropic just saw its high-stakes cybersecurity model, Mythos, accessed by unauthorized users through a third-party vendor environment. A group hunting unreleased models reportedly leveraged contractor access and guessed endpoints on the day of its announcement. Anthropic says it is investigating, with no direct breach found in its own systems.

  13. AI modelsReach 68

    Anthropic recruits Big Tech to patch the internet with unreleased Mythos model

    Anthropic just launched Project Glasswing alongside Apple, Microsoft, AWS, and Nvidia to hunt down critical software flaws using Claude Mythos Preview, an unreleased frontier model. Mythos autonomously discovered thousands of zero-days across major operating systems, including decades-old bugs in OpenBSD and FFmpeg. Anthropic is offering $100M in computing credits to defenders while keeping the model private.