{"id":991274,"date":"2026-05-29T01:31:18","date_gmt":"2026-05-29T01:31:18","guid":{"rendered":"https:\/\/www.europesays.com\/uk\/991274\/"},"modified":"2026-05-29T01:31:18","modified_gmt":"2026-05-29T01:31:18","slug":"anthropic-releases-claude-opus-4-8-promising-a-more-honest-model","status":"publish","type":"post","link":"https:\/\/www.europesays.com\/uk\/991274\/","title":{"rendered":"Anthropic releases Claude Opus 4.8, promising a more honest model"},"content":{"rendered":"

Anthropic has released<\/a> Claude Opus 4.8, an incremental upgrade to Opus 4.7. The new model is more \u2018honest\u2019 in its self-assessments and roughly four times less likely to let coding errors pass unremarked. It launches at the same price as its predecessor, alongside new features including dynamic workflows and user-controlled effort levels. Despite <\/b>assurances from Anthropic that it will eventually release a Mythos-level LLM, this doesn\u2019t appear to be anywhere close to that presumed capability.<\/b><\/strong><\/p>\n

The improvements in supposed \u201chonesty\u201d are what sets Opus 4.8 apart from its direct predecessor, according to Anthropic. The word here simply means that Claude will produce words that more closely adhere to its actions and previously implicit assessments. Early testers reported the model is quicker to flag uncertainties and less likely to make unsupported claims about its own work. According to Anthropic\u2019s alignment assessment, Opus 4.8 \u201creaches new highs on our measures of prosocial traits like supporting user autonomy and acting in the user\u2019s best interest.\u201d<\/p>\n

In terms of practical benefits, Opus 4.8 reaches new heights in agentic coding, multidisciplinary reasoning, computer use, knowledge work and financial analysis. Improvements range from less than 1 percentage point to nearly 9 percent. The difference between 4.7 and 4.8 in these statistics suggest the informal, unmeasured day-to-day experience won\u2019t be all that different at any given point, while still boosting overall results long-term. Time will tell if that is true.<\/p>\n

That aforementioned alignment assessment showed rates of deception or cooperation with misuse being substantially lower than those of Opus 4.7. Notably, those rates are now comparable to Claude Mythos Preview, the powerful model<\/a> Anthropic has kept under tight access restrictions due to its advanced cybersecurity capabilities. Pricing is unchanged for Opus, at any rate, at $5 per million input tokens and $25 per million output tokens.<\/p>\n

Dynamic workflows and effort control<\/p>\n

The launch comes bundled with several platform updates. The most significant for developers is dynamic workflows, currently available in research preview for Claude Code. It allows Claude to plan a task, then spin up hundreds of parallel subagents in a single session. Outputs are verified before being reported back. According to Anthropic, Claude Code with Opus 4.8 can now execute codebase-scale migrations across hundreds of thousands of lines of code from start to merge.<\/p>\n

Users on claude.ai also get a new effort control slider. On higher settings Claude thinks more deeply; on lower settings it responds faster and uses rate limits more slowly. The control is available on all plans. Fast mode for Opus 4.8 is now three times cheaper than it was for previous models and runs at 2.5\u00d7 the standard speed. Opus 4.7 had already introduced effort levels including an xhigh tier<\/a>, and Opus 4.8 extends and refines that foundation. These are similar features that the likes of Google and OpenAI offer to some degree, although how these sliders and controls work out will differ per vendor.<\/p>\n

The Messages API has also been updated to accept system entries inside the messages array. Developers can now update Claude\u2019s instructions mid-task without breaking the prompt cache. Overall, Anthropic really seems to have focused on usability over a flashy new LLM scorecard. Nevertheless, despite numbers being a subjective bunch, we\u2019re getting awfully close to the iterations of Claude 4.x becoming a bit overstretched. Either way, Anthropic has shuffled its choice of benchmarks around so much that no direct comparisons are yet possible from first-party results. Both benchmarks and users suggested particularly large gains made by Opus 4.6 compared to 4.5, meaning we can\u2019t really rely on the numbering to establish Anthropic\u2019s newly reached capabilities.<\/p>\n

Mythos still on its way<\/p>\n

Anthropic at least says Opus 4.8 is a \u201cmodest but tangible\u201d improvement. Larger changes are still ahead. The company is working on cheaper models with similar Opus-class capabilities, and plans to bring Mythos-class models to all customers in the coming weeks. Claude Mythos, described as a step-change above Opus, has so far only been accessible to a small group of cybersecurity organizations<\/a> as part of Project Glasswing. Anthropic says it is making swift progress on the safety safeguards required for a broader rollout.<\/p>\n","protected":false},"excerpt":{"rendered":"Anthropic has released Claude Opus 4.8, an incremental upgrade to Opus 4.7. The new model is more \u2018honest\u2019…\n","protected":false},"author":2,"featured_media":991275,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":"","_share_on_mastodon":"0"},"categories":[6],"tags":[19935,28396,51,55155,193433,275320,751,275321,16,15],"class_list":["post-991274","post","type-post","status-publish","format-standard","has-post-thumbnail","category-business","tag-ai-models","tag-anthropic","tag-business","tag-claude","tag-claude-code","tag-claude-opus-4-8","tag-generative-ai","tag-mythos","tag-uk","tag-united-kingdom"],"share_on_mastodon":{"url":"","error":""},"_links":{"self":[{"href":"https:\/\/www.europesays.com\/uk\/wp-json\/wp\/v2\/posts\/991274","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.europesays.com\/uk\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.europesays.com\/uk\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/uk\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/uk\/wp-json\/wp\/v2\/comments?post=991274"}],"version-history":[{"count":0,"href":"https:\/\/www.europesays.com\/uk\/wp-json\/wp\/v2\/posts\/991274\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/uk\/wp-json\/wp\/v2\/media\/991275"}],"wp:attachment":[{"href":"https:\/\/www.europesays.com\/uk\/wp-json\/wp\/v2\/media?parent=991274"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.europesays.com\/uk\/wp-json\/wp\/v2\/categories?post=991274"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.europesays.com\/uk\/wp-json\/wp\/v2\/tags?post=991274"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}