Anthropic launches Opus 5

Anthropic unveiled Opus 5 on Friday, the latest iteration of its flagship model. Although smaller than Fable 5, this version offers lower costs and fewer restrictions, making it the preferred choice for most applications.
Notably, Opus 5 surpasses Fable 5 across several benchmarks highlighted in the announcement.
Released just two months after Opus 4.8 (available May 28), Opus 5 follows the June launches of Mythos 5, Fable 5, and Sonnet 5. Only the lightweight Haiku model remains pending an upgrade to the 5 series.
In its announcement, Anthropic highlighted Opus 5’s improved ability to verify its work and iterate carefully until success. Benchmark tests demonstrated this capability, including instances where the model generated its own computer vision pipeline from an incomplete prompt.
Crucially, Opus 5 avoids many restrictions that have affected Fable since its release. Like its predecessor, it is exempt from the 30-day data retention policy applied to Fable and Mythos, addressing privacy concerns raised by some users.
Safeguards remain in place for cybersecurity tasks, such as exploit generation and penetration testing. For example, Opus 5 is blocked from scanning software binaries for vulnerabilities but can analyze source code, as the latter is more commonly used for defensive purposes.
Anthropic anticipates these classifiers will trigger 85% less frequently for Opus 5 than for Fable 5, reflecting a more permissive approach for the less capable model.
Additionally, Anthropic introduced a new feature to minimize disruption when safeguards activate. Users can opt into a beta feature called Automatic Fallbacks, which automatically reroutes requests to a less powerful model when a prompt triggers safety filters. This ensures API users receive a functional response rather than an error message.
Related article
Claude’s Cowork feature now recalls your past chat instructions
Anthropic’s update to Claude’s memory system removes a major friction point in using AI agents: the repetitive need to re-explain context that the AI already retains. By merging the memory architectures of standard chat and Claude Cowork, Claude now
Anthropic now has more business customers than OpenAI, according to Ramp data
Anthropic has surpassed OpenAI in verified business customers for the first time, according to this month’s AI Index from fintech firm Ramp.Data from Ramp’s client expenses reveals that 34.4% of participating businesses pay for Anthropic services, ou
Anthropic and OpenAI rivalry fades as industry shifts focus to broader AI landscape
The U.S. government is preparing to exert significant control over the release of AI models.Two weeks after the U.S. government halted Anthropic’s Fable and Mythos models, OpenAI’s latest release appears to face similar restrictions. The Information
Related Special Topic Recommendations
Comments (0)
0/500

Anthropic unveiled Opus 5 on Friday, the latest iteration of its flagship model. Although smaller than Fable 5, this version offers lower costs and fewer restrictions, making it the preferred choice for most applications.
Notably, Opus 5 surpasses Fable 5 across several benchmarks highlighted in the announcement.
Released just two months after Opus 4.8 (available May 28), Opus 5 follows the June launches of Mythos 5, Fable 5, and Sonnet 5. Only the lightweight Haiku model remains pending an upgrade to the 5 series.
In its announcement, Anthropic highlighted Opus 5’s improved ability to verify its work and iterate carefully until success. Benchmark tests demonstrated this capability, including instances where the model generated its own computer vision pipeline from an incomplete prompt.
Crucially, Opus 5 avoids many restrictions that have affected Fable since its release. Like its predecessor, it is exempt from the 30-day data retention policy applied to Fable and Mythos, addressing privacy concerns raised by some users.
Safeguards remain in place for cybersecurity tasks, such as exploit generation and penetration testing. For example, Opus 5 is blocked from scanning software binaries for vulnerabilities but can analyze source code, as the latter is more commonly used for defensive purposes.
Anthropic anticipates these classifiers will trigger 85% less frequently for Opus 5 than for Fable 5, reflecting a more permissive approach for the less capable model.
Additionally, Anthropic introduced a new feature to minimize disruption when safeguards activate. Users can opt into a beta feature called Automatic Fallbacks, which automatically reroutes requests to a less powerful model when a prompt triggers safety filters. This ensures API users receive a functional response rather than an error message.
Claude’s Cowork feature now recalls your past chat instructions
Anthropic’s update to Claude’s memory system removes a major friction point in using AI agents: the repetitive need to re-explain context that the AI already retains. By merging the memory architectures of standard chat and Claude Cowork, Claude now
Anthropic now has more business customers than OpenAI, according to Ramp data
Anthropic has surpassed OpenAI in verified business customers for the first time, according to this month’s AI Index from fintech firm Ramp.Data from Ramp’s client expenses reveals that 34.4% of participating businesses pay for Anthropic services, ou
Anthropic and OpenAI rivalry fades as industry shifts focus to broader AI landscape
The U.S. government is preparing to exert significant control over the release of AI models.Two weeks after the U.S. government halted Anthropic’s Fable and Mythos models, OpenAI’s latest release appears to face similar restrictions. The Information





Home






