frontier AI
8 stories tagged “frontier AI”.
-
Anthropic is allowing researchers from Model Evaluation and Threat Research to inspect its internal systems using employee-level access.
Anthropic gives six METR researchers employee-level access to assess AI safety and risk, with independent publ…
-
The AI race has resumed.
Google and Meta launched frontier AI models claiming parity with OpenAI and Anthropic, reopening competition p…
-
Meta's Watermelon matures into a frontier model.
Meta Superintelligence Labs launched Muse Wonder, a closed frontier AI model codenamed Watermelon, ranking sec…
-
Open models provide meaningful advantages, but do those advantages outweigh the risks?
The debate contrasts open AI models, which enable local use, research, and lower costs, with risks that powerf…
-
Google’s Gemini 4 Argon Narrows the Divide with OpenAI and Anthropic but Fails to Establish a Clear Advantage
Google Gemini 4 Argon cuts costs and improves agentic scores, but limited rollout and mixed efficiency mean no…
-
Report: Some longtime Anthropic employees are purchasing remote retreats as a hedge against "AI going wrong"
According to The Wall Street Journal, several longtime Anthropic employees are considering purchasing remote U…
-
Trump’s AI risk strategy relies on Big Tech allies to self-regulate
Trump's AI risk approach uses voluntary Big Tech commitments, independent audits, and common benchmarks to man…
-
Florida cites extinction risks in legal effort to stop OpenAI development
Florida seeks an OpenAI injunction, citing extinction fears, utility risks, false safety claims, child access,…