Search

61 article(s)

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