Anthropic's Claude Opus 4.8 is four times more honest, Mythos next
Anthropic has released Claude Opus 4.8, an upgrade to its flagship AI model that is four times less likely to let code flaws pass unremarked. The company also teased Mythos-class models, which have already found more than 10,000 critical software vulnerabilities through Project Glasswing, and announced a $65 billion Series H round at a $965 billion post-money valuation.
Anthropic’s Claude Opus 4.8 is four times more honest, Mythos next Anthropic has launched Claude Opus 4.8, an upgraded AI model emphasizing increased honesty, reliability in agentic tasks, and improved self-correction, available at the same price as its predecessor. The new model shows marked improvements across benchmarks and alignment assessments, with early adopters reporting significant benefits in coding, legal, and finance applications. Alongside Opus 4.8, Anthropic is rolling out new features like effort control and dynamic workflows, and has teased upcoming Mythos-class models with higher intelligence, which have already been instrumental in identifying thousands of software vulnerabilities.
- Claude Opus 4.8, Anthropic’s latest AI model, is now available, offering improvements in honesty, reliability, and self-correction.
- The model shows higher scores on benchmarks for agentic coding, reasoning with tools, computer use, and knowledge work.
- Opus 4.8 demonstrates higher prosocial traits and reduced misaligned behaviors compared to its predecessor.
- Endorsements from companies like Cognition, Cursor, Harvey, Databricks, Thomson Reuters, and Hebbia highlight Opus 4.8’s performance enhancements.
- New features include effort control on claude.ai and Cowork, and dynamic workflows for Claude Code.
- Anthropic is preparing to release Mythos-class models, described as having higher intelligence and already used in cybersecurity to find over 10,000 vulnerabilities.
- The company secured a $65 billion Series H funding round at a $965 billion valuation, alongside international expansion.
- Opus 4.8 focuses on reliability, aiming to be more useful in agentic workflows with limited human oversight, positioning Anthropic in a competitive AI market.
- The Messages API has been updated to accept system entries within the messages array for easier instruction updates.
- A faster mode for Opus 4.8 is now available and is cheaper than for previous models. Continue reading https://foxvector.com/articles/f735f9af-a529-4bdb-ac99-77b1196474ef
Write a comment