Editorial illustration for Claude Opus 4.6 adds 1M-token context, teams; used by 44% of enterprises
Claude Opus 4.5: Million-Token AI for Enterprise
Claude Opus 4.6 adds 1M-token context, teams; used by 44% of enterprises
Anthropic just rolled out Claude Opus 4.6, a model that stretches its context window to a full million tokens and introduces “agent teams” designed to tackle complex coding tasks. While the tech is impressive on paper, its real test lies in how quickly businesses put it to work. The upgrade arrives as OpenAI’s Codex continues to dominate developer pipelines, prompting Anthropic to push a desktop‑focused version that directly challenges Claude Code’s growing traction.
Early reports suggest the new capabilities have already reshaped software‑development workflows, especially since the latter half of 2024. In that climate, enterprises are weighing whether the expanded context and coordinated agents can deliver tangible productivity gains. The numbers coming out of Anthropic’s own rollout hint at a shift in adoption patterns, and the company is touting a benchmark win that could tip the scales further.
---
*Forty‑four percent of enterprises now use Anthropic in production, driven by rapid capability gains in software development since late 2024. The desktop launch is a strategic counter to Claude Code’s momentum. According to Anthropic’s announcement, Opus 4.6 achieves the highest score on Terminal‑Ben.*
Forty-four percent of enterprises now use Anthropic in production, driven by rapid capability gains in software development since late 2024. The desktop launch is a strategic counter to Claude Code's momentum. According to Anthropic's announcement, Opus 4.6 achieves the highest score on Terminal-Bench 2.0, an agentic coding evaluation, and leads all other frontier models on Humanity's Last Exam, a complex multi-discipline reasoning test. On GDPval-AA -- a benchmark measuring performance on economically valuable knowledge work tasks in finance, legal and other domains -- Opus 4.6 outperforms OpenAI's GPT-5.2 by approximately 144 ELO points, which translates to obtaining a higher score approximately 70% of the time.
What does Claude Opus 4.6 really deliver? The model now supports a 1 million‑token context window and introduces “agent teams,” features Anthropic says let it sustain longer autonomous workflows and plan more carefully. In Anthropic’s own benchmarks the upgrade outperforms OpenAI’s GPT‑5.2 on key enterprise tests, and it claims the highest score on Terminal‑B.
Forty‑four percent of enterprises already run Anthropic in production, a share the company attributes to rapid capability gains in software development since late 2024. OpenAI’s response—a Codex desktop application launched three days later—signals a direct challenge to Claude’s momentum. Yet the competitive edge remains uncertain; the article offers no data on real‑world performance beyond internal benchmarks, nor does it explain how “agent teams” differ in practice from existing tooling.
The timing is notable, arriving amid broader turbulence in the AI sector and software markets, but whether the new context length and team‑based agents translate into lasting enterprise advantage is still unclear.
Further Reading
- Latest Claude Opus 4.6 Model Upgrade: 1M Token Context and Enhanced Reliability Analysis - Blockchain News
- Models overview - Claude API Docs - Anthropic Claude API Documentation
- Claude Opus 4.6 System Card - Anthropic
- Anthropic says Claude Opus 4.6 supports a 1M context window, scored 90.2% on BigLaw Bench - Techmeme
- Claude Opus 4.6 - Hacker News - Hacker News
Related Reading
Browse all Anthropic news →
Business & Startups OpenAI, a Series F San Francisco startup founded in 2015 by eight pioneers
Business & Startups Terminal-Bench 2.0 launches with Harbor, testing any container-installable agent
Business & Startups Zuckerberg Unveils Meta Compute to Build Global AI Infrastructure
Related Topic Anthropic's Claude also citing Elon Musk's Grokipedia, reports say
LLMs & Generative AI
Related Topic Claude Code Update Boosts Context Window for Complex Developer Workflows
LLMs & Generative AI
![OpenAI logo with spiraling pastel colors, representing the launch of their Frontier platform for AI agents [techcrunch.com].](https://cms.aidailypost.com/uploads/openai_launches_frontier_platform_build_deploy_manage_ai_agents_a4a42ba243.webp)
From Archives OpenAI launches Frontier platform to build, deploy, manage AI agents
From Archives Alphabet posts USD 400 B revenue, YouTube tops streaming, 325 M paid subs
From Archives Sam Altman says OpenAI’s Super Bowl ad focuses on builders, not Anthropic jokes
From Archives GitHub adds Claude and Codex agents to Copilot’s Agent HQ platform
LLMs & Generative AI
Common Questions Answered
How does Claude Opus 4.5 improve coding capabilities compared to previous versions?
[allthings.how](https://allthings.how/claude-opus-4-5-is-anthropics-new-flagship-for-coding-and-agents/) reports that Opus 4.5 achieves an 80.9% score on the SWE-bench Verified evaluation for software bug-fixing, which is ahead of previous Claude versions. The model is designed to handle advanced coding tasks, including multi-day software projects compressed into hours, with improvements across programming languages, project planning, and software architecture.
What are the key pricing and availability details for Claude Opus 4.5?
According to [allthings.how](https://allthings.how/claude-opus-4-5-is-anthropics-new-flagship-for-coding-and-agents/), Claude Opus 4.5 is available through multiple platforms including Anthropic's own apps, Claude API, Amazon Bedrock, Google Cloud's Vertex AI, and Microsoft Foundry. The pricing is set at $5 per 1M input tokens and $25 per 1M output tokens, with Anthropic removing Opus-specific usage caps for users with Opus 4.5 access.
What unique technological innovations does Claude Opus 4.5 introduce?
[applyingai.com](https://applyingai.com/2025/11/anthropics-opus-4-5-elevating-claudes-coding-and-agentic-prowess-in-the-ai-landscape/) highlights two key innovations: a Modular Code-Optimized Transformer with specialized code-focused attention layers and Vectorized Execution Traces. These advancements allow the model to adaptively shift parameters when recognizing programming syntax and improve token prediction accuracy in complex code blocks, making it more effective for software engineering tasks.