Search

37 article(s)

Two autonomous systems achieve gold-level performance on third-party-graded tasks, OpenAI prepares a misalignment disclosure framework, GPT-6 Astra arrives everywhere

Two autonomous systems achieve gold-level performance on third-party-graded tasks, OpenAI prepares a misalignment disclosure framework, GPT-6 Astra arrives everywhere

September 5, 2026: Meta’s autonomous research system and an NVIDIA-tuned Nemotron achieve gold-level performance on third-party-graded tasks, OpenAI announces a framework for reporting misalignment incidents, and GPT-6 Astra arrives in v0, Perplexity Computer, Codex CLI, and Genspark.

metallamaopen-source +25
Claude Fable 5.1 and Mythos 5.1, a single model with two guardrail levels, OpenAI classifies Astra at the Critical cybersecurity threshold, Perplexity splits a task between the cloud and the Mac

Claude Fable 5.1 and Mythos 5.1, a single model with two guardrail levels, OpenAI classifies Astra at the Critical cybersecurity threshold, Perplexity splits a task between the cloud and the Mac

September 1, 2026: Anthropic releases Claude Fable 5.1 and Claude Mythos 5.1, a single model offered with two guardrail levels and adopted the same day by seven coding tools. OpenAI classifies its upcoming Astra model at the Critical cybersecurity threshold, while Perplexity runs a single task across the cloud and a local model on Mac.

claudeanthropicclaude-code +32