7/30/2026, 1:03:57 PM · foundation-models

Anthropic Releases Claude Opus 5 at Half the Cost of Its Frontier Fable Tier, Joining a Crowded July Model Cycle

Anthropic's Claude Opus 5 launched July 24, 2026, offering near-frontier performance for enterprise and developer workloads at the same price point as its predecessor but roughly half the cost of the company's top-tier Fable 5, entering a market that saw GPT-5.6, Grok 4.5, and Meta Muse Spark 1.1 all ship within weeks.

Overview

Anthropic released Claude Opus 5 on July 24, 2026, positioning it as the company's workhorse model for enterprise users, knowledge workers, and developers. <cite index="2-6">Anthropic says the model delivers nearly all the intelligence of its top-of-the-line Claude Fable 5 at half the cost.</cite> <cite index="2-7">The model is priced at $5 per million input tokens and $25 per million output tokens — unchanged from its predecessor, Opus 4.8.</cite>

Model Capabilities

<cite index="1-1">Claude Opus 5 ships with a one-million-token context window and an "xhigh" reasoning effort mode.</cite> <cite index="3-11,3-12">Anthropic gives users an effort "dial" that lets them decide how much computing power the model should devote to a task, with lower effort levels preserving much of the model's performance while using fewer tokens and costing less to operate.</cite> <cite index="8-5,8-6">The company also says the model is easier to use than its predecessors — it verifies its own work and recovers from errors without human intervention.</cite>

<cite index="4-8">According to Anthropic, Opus 5 is particularly efficient on software engineering and knowledge work tasks, achieving leading results on benchmarks including Frontier-Bench and GDPval-AA.</cite> <cite index="2-3">On scientific research, Opus 5 scores 10.2 percentage points higher than Opus 4.8 on Anthropic's internal chemistry benchmark, though the company acknowledged that its Mythos 5 model remains stronger for long-horizon, open-ended work such as autonomous drug design campaigns.</cite>

<cite index="3-7">Anthropic said Opus 5 is "the most aligned Opus model and is the least susceptible to being tricked into misuse."</cite> The company also noted that government safety review work continues: <cite index="3-9,3-10">"We continue to work with our government partners to conduct their own independent testing of our models. This includes Opus 5," an Anthropic spokesperson told Axios.</cite>

Distribution and Tier Placement

<cite index="2-8">Opus 5 becomes the new default model on Claude Max, Anthropic's premium consumer tier, and the strongest model available on Claude Pro.</cite> <cite index="3-13">Anthropic also launched an option for users to switch which model they are using mid-task, which can be used to keep artificial intelligence (AI) costs in check.</cite>

<cite index="8-7">Opus 5 is the fourth model Anthropic has released in less than two months, following Mythos 5, Fable 5, and Sonnet 5 in June.</cite>

Competitive Context

The Opus 5 release lands inside what observers have called an unusually compressed frontier model cycle. <cite index="9-1">GPT-5.6 (Generative Pre-trained Transformer 5.6), a large language model (LLM) developed by OpenAI, was publicly released on July 9, 2026.</cite> <cite index="9-2,9-3">It comes in three variants — Luna, Terra, and Sol — designed to expand user capabilities across enterprise work, coding, scientific research, and cybersecurity.</cite> <cite index="9-7">Due to government restrictions levied by the Trump administration, GPT-5.6 was first released as a limited preview only on June 26, 2026 for a "small group of trusted partners."</cite>

SpaceXAI's Grok also entered its latest iteration during the same window. <cite index="21-14">Grok 4.5, SpaceXAI's model for coding, agentic tasks, and knowledge work, became available on the xAI application programming interface (API).</cite> <cite index="19-1">The public announcement in early July 2026 marked Grok 4.5 as a V9-based model tuned on real developer activity and aimed at coding, agents, and knowledge work, rather than conversational use alone.</cite>

Meta added a third entrant to the July cycle. <cite index="29-3,29-4">On July 9, 2026, Meta introduced Muse Spark 1.1, the latest model from Meta Superintelligence Labs — a multimodal reasoning model built for agentic tasks, with major gains in tool and computer use, coding, and multimodal understanding.</cite> <cite index="27-2">The release included a new paid tier for developers, marking the first time Meta has charged businesses for access to its models.</cite> <cite index="26-8">API pricing was set at $1.25 per million input tokens and $4.25 per million output tokens.</cite>

Commercial Context

<cite index="2-5">Reuters reported in February that Anthropic was valued at roughly $380 billion in its latest funding round, following a period in which its annualized revenue climbed from about $1 billion at the end of 2024 to a projected $9 billion by the end of 2025, with internal targets reportedly reaching $20 to $26 billion for 2026.</cite> <cite index="4-12,4-13">Separately, Anthropic recently announced plans to use up to 2GW of AMD Instinct MI450 Series graphics processing units (GPUs) in AMD Helios rack-scale systems, alongside a strategic equity investment of up to $5 billion from AMD.</cite>

Cross-references

Sources

  1. [1]
    What Is Claude Opus 5? Anthropic's Honeycomb Flagship
  2. [2]
    Anthropic launches Claude Opus 5, a cheaper AI model for coding, agents and enterprise workflows | VentureBeat
  3. [3]
    Anthropic releases new model, Opus 5
  4. [4]
    Anthropic releases Claude Opus 5 AI model on all platforms
  5. [5]
    Claude (language model)
  6. [6]
    Anthropic Launches Claude Opus 5
  7. [7]
    Anthropic Claude Model Release Timeline - Model Family Tree, Capability Evolution, and Platform Availability | hidekazu-konishi.com
  8. [8]
    Anthropic releases Claude Opus 5: Here's how it's different than what's already out there | Fortune
  9. [9]
    GPT-5.6 - Wikipedia
  10. [10]
    OpenAI to release GPT-5.6 Sol, Terra and Luna on July 9 - Neowin
  11. [11]
    GPT-5.6 Release Date: When You Can Actually Use It
  12. [12]
    GPT-5
  13. [13]
    GPT-4.5
  14. [14]
    GPT-5.4
  15. [15]
    GPT-5.5
  16. [16]
    GPT-5.2
  17. [17]
    OpenAI Rumored to Launch GPT-5.6 to Public, Exactly Timing Anthropic Fable 5 Limit Expiration
  18. [18]
    Grok 4.5 released: Architecture, release date, and other we know - CometAPI
  19. [19]
    Grok 4.5: How xAI’s New Model Competes on Code and Cost
  20. [20]
    Grok (chatbot)
  21. [21]
    Release Notes - xAI Docs - SpaceXAI
  22. [22]
    SpaceXAI — Creators of Grok, the AI Chatbot
  23. [23]
    xAI's Grok Roadmap: 7 Models in Training Now, Grok 5 at 10 Trillion Parameters — Full Timeline | MindStudio
  24. [24]
    Grok 4.5, 4.6, and Tavily: 4 Updates That Matter
  25. [25]
    Grok (JPEG 2000)
  26. [26]
    Muse Spark 1.1: Meta's Agentic Model and API | DataCamp
  27. [27]
    Meta Starts Charging for AI With Muse Spark 1.1 Agentic Model - Bloomberg
  28. [28]
    Meta enters the crowded AI coding battle with Muse Spark 1.1 | TechCrunch
  29. [29]
    Introducing Muse Spark 1.1
  30. [30]
    Meta prices Muse Spark 1.1 API at $1.25/$4.25 per M tokens | AI Weekly
  31. [31]
    Meta AI Doesn’t Just Think, It Acts
  32. [32]
    Muse Spark 1.1 — Meta Model API (July 2026)
  33. [33]
    Meta releases latest update of AI model Muse Spark as tech giant accelerates AI push under Alexandr Wang | Fortune