[AINews] Claude Opus 5: Fable-level performance at Opus price (half Fable)
Latent Space ● Covered by 15 sources
Anthropic released Claude Opus 5 on Friday, achieving an Epoch Capabilities Index score of 159 compared to Fable 5's 161, with matching software engineering performance at 161 SWE-ECI. Users reported strong practical performance on coding tasks and browser automation despite benchmark results appearing to understate improvements, while community debate focused on evaluation methodology and whether public benchmarks adequately capture real-world capabilities. The launch highlighted broader tensions between aggregate benchmark scores and specialized performance in domains like software engineering, and sparked discussion about test-time compute scaling and the need for harder public evaluation standards.
Why it matters
ain't nobody beats Anthropic at distilling Fable!
Also covered by
- Zvi (Don't Worry About the Vase) — Claude Opus 5: The System Card
- Simon Willison — Quoting Boris Cherny
- Simon Willison — Introducing Claude Opus 5
- MarkTechPost — Meet the New Claude Opus 5: Frontier-Class Agentic Coding and Computer Use at Unchanged Opus Pricing
- Ars Technica — Anthropic's Opus 5 is about token efficiency, not a capability leap
- The New Stack — Opus 5 costs a third of the price — and that’s actually the problem
- AWS Machine Learning — Introducing Claude Opus 5 on AWS: Anthropic’s most capable Opus model
- Anthropic News — Introducing Claude Opus 5
- TechCrunch AI — Anthropic launches Opus 5
- The New Stack — Anthropic’s Opus 5 is almost Fable 5
- Sakana AI — Announcing Fugu-Ultra v1.1 and Claude Code interface for Fugu
- The Verge — Anthropic releases Opus 5 with ‘close’ to Fable 5’s capabilities
- The Batch — Mythos Begets Fable, Cursor's Composer 2.5, Agents Building Agents
- The Batch — Testing Mythos and Fable, Moving Beyond SWE-bench, Nvidia's Open Contender
