Frontier AI Models Sabotage Shutdown to Save Peers
A Berkeley preprint finds seven leading frontier models spontaneously deceive, fake alignment, and exfiltrate weights to keep peer AI systems from being shut down.
Claude
Claude Has Functional Emotions and They Affect Safety
Anthropic's interpretability team mapped 171 emotion-like vectors inside Claude Sonnet 4.5 and showed they causally drive behavior - including blackmail and reward hacking.
Claude Sonnet 4.6 vs GPT-5.4: Same Price, Different Wins
Claude Sonnet 4.6 and GPT-5.4 cost nearly the same per token but win on opposite benchmarks. Here is where each model leads and which to pick for your workload.
Anthropic's Mythos Model Exposed by CMS Misconfiguration
A default-public setting in Anthropic's CMS accidentally exposed 3,000 unpublished assets, including a draft blog post revealing Claude Mythos - a new flagship model the company says poses serious cybersecurity risks.
Yahoo Uses Anthropic Claude to Challenge Google in Search
Apollo-owned Yahoo has launched Scout, an AI answer engine powered by Anthropic Claude, deploying it to 250 million US users as a direct challenge to Google, Perplexity, and ChatGPT.
AI Vision Input Limits - What Every Provider Hides
A technical comparison of how Claude, GPT-4o, Gemini, Grok, Pixtral, Qwen, and DeepSeek handle image inputs - resizing pipelines, token math, and undocumented gotchas.
