Anthropic Releases New AI Model, Claude Opus 4.8 Anthropic is pitching its latest flagship AI, Claude Opus 4.8, as both a technical leap and a character upgrade—an AI that not only codes better, but is more willing to admit when it’s unsure.

On May 28, Anthropic formally unveiled Opus 4.8 as an upgrade over Opus 4.7, keeping the same price while promising broader benchmark gains and a “more effective collaborator.” The company highlighted new features: user “effort” controls in the Claude interface, a faster and cheaper “fast mode,” and a “dynamic workflows” system in Claude Code that can orchestrate hundreds of parallel sub‑agents on large problems.

Early coverage framed the model around honesty and reliability. The Verge reported that Anthropic trains “all its models to be honest” and that testers found Opus 4.8 “more likely to flag uncertainties about its work and less likely to make unsupported claims,” estimating it is “around 4x less likely than its predecessor to allow flaws in code it’s written to pass unremarked.” The Next Web similarly called it “more honest, more reliable in agentic tasks, and better at catching its own mistakes,” noting price parity with Opus 4.7.

Tech outlets also placed the launch in a competitive timeline. TechCrunch observed that Opus 4.8 arrived just 41 days after Opus 4.7, a “much faster upgrade cycle than normal,” amid pressure from OpenAI and Google. Axios stressed performance gains in “agentic coding, reasoning, financial analysis and knowledge work,” plus a fast mode that runs at 2.5x speed and is three times cheaper than previous models.

Independent reviewers focused on real‑world feel. Every’s “vibe check” called Opus 4.8 “a legitimately great model” that “bests GPT-5.5 on our Senior Engineer benchmark” and is “the best model we’ve tested for writing and knowledge work,” arguing Anthropic “could have called this Opus 5 and none of us would have blinked.” But they criticized the Claude app as “a mess,” saying the harness of Chat, Code, and Cowork tabs makes it “hard-to-use” despite the model’s strength.

Meanwhile, Anthropic and the press framed Opus 4.8 as a precursor to Mythos, a more powerful line still constrained by security concerns. Axios reported that Mythos‑class models are expected “in the coming weeks,” while The Next Web noted Mythos has already helped find over 10,000 critical software vulnerabilities.

Online reaction ranged from technical benchmarks to memes. A viral clip joking about using Opus 4.8 just to rename a file drew a simple “😂” from Elon Musk on X, signaling both curiosity and skepticism from a high‑profile AI rival.

Continue reading https://foxvector.com/stories/019e79e0-17d0-3f92-70c3-321ab7e11259