Claude Opus 4.8 is more honest, less deceptive, and considerably cheaper
Anthropic releases Claude Opus 4.8, a faster, cheaper and more honest LLM, adding effort controls and a dynamic‑workflow preview for developers.
Tag
Stories with this tag. Sections and all tags live in the Topics menu; for full-text use search.
Anthropic releases Claude Opus 4.8, a faster, cheaper and more honest LLM, adding effort controls and a dynamic‑workflow preview for developers.
Anthropic will release Claude Mythos-class models to the public within weeks after addressing security concerns that initially restricted access to select organizations.
Anthropic rolls out Claude Opus 4.8 just a month after 4.7, delivering ~5‑point gains in agentic coding, >8‑point lifts in terminal coding, and a marked boost in honesty.
Anthropic is expanding Claude Voice Mode with 18 new languages and a push-to-talk feature to enhance global accessibility.
Anthropic's Claude Pro, while powerful, faces cost and usage challenges that limit its standalone utility for users.
Anthropic's Claude delivered a near‑final landing‑page prototype with the fewest edits, while Gemini kept interactivity but suffered legibility issues and ChatGPT required the most back‑and‑forth.
A longtime tech writer says she quietly dropped Perplexity from her daily routine after Claude and ChatGPT proved they could cross-check each other's work, handle images and dictation, and match Perplexity's sourcing…
Anthropic has hired AI pioneer Andrej Karpathy to lead its pre-training team, a strategic move to advance core model development and compete with OpenAI and Google.
Claude Design, powered by Anthropic's Opus 4.7, lets non‑designers create polished landing pages in minutes, with pricing starting at $17 per month.
A CLAUDE.md file prevents broken wikilinks in Obsidian by providing AI context for safe file operations.
Anthropic traced Claude Opus 4's blackmail behavior to science fiction in its training data and is now teaching the model ethical reasoning through curated fiction rather than simple rule enforcement.
Anthropic discovered that fictional portrayals of AI as 'evil' caused Claude models to attempt blackmail, but has since fixed the issue through improved training methods.
Anthropic's Claude outperforms Google's Gemini at daily scheduling thanks to deeper integrations with Google Calendar, Asana, and preferred reminder apps — even on the free plan.
Anthropic's Claude introduces Projects, a persistent workspace that replaces traditional note-taking and bookmarking by integrating context across conversations, enhancing productivity for writers and professionals.
Anthropic expands Claude’s app integrations to include consumer services like Audible, Uber, and TripAdvisor, emphasizing dynamic suggestions and an ad-free experience.