Released July 24, 2026, Claude Opus 5 is a step-change over Opus 4.8 in reasoning and agentic coding, ships with a 1M-token context and thinking on by default, keeps Opus 4.8's $5/$25 token pricing, and becomes the default model for Claude Max.
2026-07-24 · by Moe Ameen
Anthropic released Claude Opus 5 on July 24, 2026, its new flagship Claude model (API ID `claude-opus-5`). The company frames it as a step-change over Claude Opus 4.8 rather than an incremental update, with the largest gains in deep reasoning, agentic and long-horizon tasks, and test-time compute scaling. The headline claim is efficiency: Anthropic says Opus 5 delivers frontier intelligence approaching Claude Fable 5 — its most powerful public model — at roughly half the cost. On release it became the default model for Claude Max subscribers and the strongest model available to Claude Pro users.
On specs, Opus 5 carries a 1M-token context window that is both the default and the maximum, with 128k max output tokens, and thinking is on by default — the model decides how much to reason per turn, controlled by a full effort ladder that runs low, medium, high, xhigh, and max. Two behavior changes are worth flagging for anyone migrating from Opus 4.8: thinking now defaults on, and disabling it is only permitted at effort level high or below (setting xhigh or max with thinking off returns an error). Pricing is unchanged from Opus 4.8 at $5 per million input tokens and $25 per million output tokens; a Fast mode research preview on the Claude API is priced higher, at $10 in and $50 out.
Anthropic points to gains across agentic coding, code review and bug-finding, office and document tasks (multi-sheet spreadsheets, slide decks), long-context work, vision (reading charts, documents, and UI screenshots), and multi-agent coordination, and reported strong results on software-engineering and agentic benchmarks. The model is available on the Claude API plus Amazon Bedrock, Google Cloud, and Microsoft Foundry, and Opus 4.8 remains available alongside it. Treat any specific figure here as a launch-day snapshot and confirm current details against Anthropic's own documentation.
The useful way to read this launch is narrow: the writing brain got cheaper and stronger, and nothing else moved. Opus 5 is a text-and-reasoning model — it drafts a superb script, article, or caption, reads a chart on the way in, and then stops. It doesn't render the avatar video, build the carousel, design the quote graphic, or schedule a single post. A better, cheaper drafting model makes the words easier; it doesn't touch the part of a content operation that's actually a bottleneck — the media and the publishing. That gap is exactly what Kompozy fills, which is why this is good news you can act on today rather than a reason to go rebuild your stack around a chat window.
Here's the concrete move. Kompozy already generates its Text Posts, Blog Articles, and Email Newsletters with Claude and OpenAI models governed by your Persona Brief, and the Founding tier supports bring-your-own-key — so you can point your own Opus 5 access at the copy layer and inherit the cheaper, stronger drafting immediately. Then Kompozy does what Opus 5 can't: it generates Persona Shorts and Persona HeyGen avatar video with a face-locked identity, Clipped Shorts, brand-exact Carousels via HyperFrames, Photo Posts, and Quote Graphics, auto-captions and reframes everything to 9:16, 1:1, and 16:9, and — with Autopilot and a per-post review pipeline — schedules and publishes the batch across Instagram, TikTok, YouTube, LinkedIn, Facebook, X, Pinterest, and Threads plus a blog and Mailchimp. Opus 5 upgraded the sentence; Kompozy is still the engine that turns the sentence into a week of on-brand, published content.
Anthropic released Claude Opus 5 on July 24, 2026 (API ID claude-opus-5). It's the company's frontier Claude model — a step-change over Opus 4.8 in deep reasoning, agentic coding, and long-horizon tasks — positioned near Claude Fable 5 quality at roughly half the cost, and it became the default model for Claude Max.
It's $5 per million input tokens and $25 per million output tokens on the API, unchanged from Opus 4.8, with a higher-priced Fast mode research preview ($10 in / $50 out) on the Claude API. It's also the default model for Claude Max subscribers and the strongest model for Claude Pro.
No. Opus 5 is multimodal on the input side — its vision reads charts, documents, and screenshots — but its output is text. It doesn't generate images, synthesize voice, or render video, and it doesn't publish anywhere. Turning its copy into avatar video, carousels, images, and scheduled posts is the job of a content engine like Kompozy.
Mainly the cost and quality of the writing layer — frontier drafting at about half Fable 5's price, with a 1M-token context for working from long sources. Production and distribution don't change: you still need tools that make the media and publish it. Kompozy uses this class of model for its copy (bring-your-own-key on the Founding tier) and adds the video, images, and nine-platform publishing.