Anthropic unveils Opus 4.5 with major performance and recollection upgrades
Anthropic has introduced Opus 4.5, the recent and ultimate free in its 4.5 series of AI models. The rollout follows earlier launches of Sonnet 4.5 in
An illustrative image related to: Anthropic unveils Opus 4.5 with major performance and recollection upgrades, highlighting key aspects of the story. | Image source: The Express Tribune
Anthropic unveils Opus 4.5 with major performance and recollection upgrades
Anthropic has introduced Opus 4.5, the recent and ultimate free in its 4.5 series of AI models. The rollout follows earlier launches of Sonnet 4.5 in September and Haiku 4.5 in October.
The firm says Opus 4.5 delivers state-of-the-art results across several benchmark categories, including coding tasks (SWE-Bench, Terminal-bench), tool-apply evaluations (tau2-bench, MCP Atlas), and general reasoning tests such as ARC-AGI 2 and GPQA Diamond. It is plus the first pattern to score above 80% on SWE-Bench Verified, a widely cited benchmark for script generation accuracy.
Anthropic is pairing the pattern with upgrades to its productivity tools. With the free of Opus 4.5, the firm is expanding reach to Claude for Chrome and Claude for Excel. The Chrome extension will now be available to all Max users, while the Excel-focused tool will be offered to Max, Squad, and Enterprise subscribers.
The pattern includes significant improvements in long-setting performance, requiring changes to how it handles and manages recollection. Speaking to TechCrunch , Anthropic’s head of product management for research, Dianne Na Penn, said long setting alone is not enough for high-standard result. “Knowing the right details to remember is really crucial in addition to just having a longer setting window,” she said.
Those changes have enabled a long-requested “endless chat” trait for paying Claude users. Instead of stopping when the pattern reaches its setting boundary, Opus 4.5 can now condense earlier parts of the conversation automatically, without notifying the user.
Anthropic says many of the upgrades are designed for complex “agentic” workflows in which Opus serves as a guide framework orchestrating tasks handled by smaller, Haiku-powered agents. These scenarios need strong working-recollection capabilities, an sector the firm claims has seen substantial gains in this free.
Source & Editorial Note
Curious Pakistan provides a concise, contextual summary of verified newsroom reports and official statements.
Disclaimer: This article contains an editorial summary of news reported by third-party publishers. Curious Pakistan does not claim original reporting; information is adapted only for clarity, context and readability
Contact Us For More Detail Or any Query
{"@context":"https://schema.org","@type":"NewsArticle","headline":"Anthropic unveils Opus 4.5 with major performance and recollection upgrades","image":["https://images.weserv.nl/?url=i.tribune.com.pk%2Fmedia%2Fimages%2F2colm-161750893360-0%2F2colm-161750893360-0-640x480.webp&w=1200&h=675&fit=cover&output=webp"],"datePublished":"2026-03-29","dateModified":"2026-03-29","author":[{"@type":"Person","name":"Insight Desk","url":"https://www.curiouspakistan.com/p/author.html?label=Z-AI"}],"publisher":{"@type":"Organization","name":"Curious Pakistan","logo":{"@type":"ImageObject","url":"https://blogger.googleusercontent.com/img/a/AVvXsEgL5seAo-05Wvrnele_PZkLSSiSNLr4ljkDQfTwjQtHnEYu6fDfDqUR5HOnqcWoLe-O6JuW0ZNnc5mVBjQewHUJhKTLtHQ1ouwj3Ge5U_TrTQ0wQj-HnETxIONhcECufKT71ikUZV3-14RLOBldvj8Nu8KR8WQSMwKGdNaaa7FqmWYLC6tmDg8mVD2fqw=s725"}},"description":"Anthropic has introduced Opus 4.5, the recent and ultimate free in its 4.5 series of AI models. The rollout follows earlier launches of Sonnet 4.5 in September and Haiku 4.5 in October. The firm says Opus 4.5 delivers state-of-the-art results across several benchmark categories, including coding","articleBody":"Anthropic has introduced Opus 4.5, the recent and ultimate free in its 4.5 series of AI models. The rollout follows earlier launches of Sonnet 4.5 in September and Haiku 4.5 in October.\\n\\nThe firm says Opus 4.5 delivers state-of-the-art results across several benchmark categories, including coding tasks (SWE-Bench, Terminal-bench), tool-apply evaluations (tau2-bench, MCP Atlas), and general reasoning tests such as ARC-AGI 2 and GPQA Diamond. It is plus the first pattern to score above 80% on SWE-Bench Verified, a widely cited benchmark for script generation accuracy.\\n\\nAnthropic is pairing the pattern with upgrades to its productivity tools. With the free of Opus 4.5, the firm is expanding reach to Claude for Chrome and Claude for Excel. The Chrome extension will now be available to all Max users, while the Excel-focused tool will be offered to Max, Squad, and Enterprise subscribers.\\n\\nThe pattern includes significant improvements in long-setting performance, requiring changes to how it handles and manages recollection. Speaking to TechCrunch , Anthropic’s head of product management for research, Dianne Na Penn, said long setting alone is not enough for high-standard result. “Knowing the right details to remember is really crucial in addition to just having a longer setting window,” she said.\\n\\nThose changes have enabled a long-requested “endless chat” trait for paying Claude users. Instead of stopping when the pattern reaches its setting boundary, Opus 4.5 can now condense earlier parts of the conversation automatically, without notifying the user.\\n\\nAnthropic says many of the upgrades are designed for complex “agentic” workflows in which Opus serves as a guide framework orchestrating tasks handled by smaller, Haiku-powered agents. These scenarios need strong working-recollection capabilities, an sector the firm claims has seen substantial gains in this free.","articleSection":"Technology News","keywords":["Breaking News","Technology News"]}
