The one AI update too big to ignore this week

The one AI update too big to ignore this week.

Anthropic just shipped its third major model in two months. But Claude Opus 4.5 isn't just another benchmark bragging session. It's Anthropic saying: we're done being the "safe but expensive" option.

📌 THE RUNDOWN

Anthropic's Opus 4.5 Rewrites the Rules on Price and Power

Claude Opus 4.5 launched November 24, 2025. Anthropic calls it "the best model in the world for coding, agents, and computer use."

➡️ The move: Opus 4.5 achieved 80.9% on SWE-bench Verified, the first model to cross the 80% threshold. It beat OpenAI's GPT-5.1-Codex-Max, Google's Gemini 3 Pro, and Anthropic's own Sonnet 4.5 on real-world software engineering tasks. Anthropic tested it on their notoriously difficult take-home exam for performance engineering candidates. Within the two hour time limit, Opus 4.5 scored higher than any human candidate ever had.

But here's where it gets interesting for your budget: pricing is $5 per million input tokens and $25 per million output tokens, making Opus-level intelligence accessible at one-third the cost of previous offerings. The previous Opus ran at $15/$75. That's a 67% price cut on the smartest model they've ever shipped.

⚡Why it matters: Anthropic is attacking from both ends. They dropped the price floor while raising the capability ceiling. At medium effort, Opus 4.5 matches Sonnet 4.5's best benchmark score while using 76% fewer output tokens. Translation: you can run the flagship model for roughly what the mid-tier cost before.

This isn't generosity. OpenAI's GPT-5.1 arrived just days earlier. Google's Gemini 3 Pro dropped the week before. Anthropic needed a counterpunch that competed on both intelligence and economics. They delivered.

🎯 Your Takeaway: Anthropic's real upgrade isn't the model. It's finally pricing like they want market share.

Infinite Chat Solves the "Start Over" Problem

Infinite Chat is a new feature available to paid subscribers that fixes one of the top complaints from Claude users: context window limit errors.

➡️ The move: Rather than simply increasing the token ceiling, the system avoids context-window failures altogether by compacting, indexing, and retrieving prior states. Earlier parts of a conversation are summarized and re-encoded so that logically important constraints are preserved even when the raw text has been pushed far into the past.

"Within Claude AI, within the product itself, you effectively get this kind of infinite context window due to the compaction, plus some memory things that we're doing," explained Alex Albert, Anthropic's head of developer relations.

⚡Why it matters: Context limits have been the quiet tax on long AI sessions. Hit the wall mid-project, and you're re-pasting background, re-explaining context, losing momentum. From a user's perspective, multi-day or multi-week interactions feel continuous. A long-running refactor, a book-length legal analysis, or a translation project extending over hundreds of thousands of words can remain coherent without the user repeatedly re-pasting background material.

🎯 Your Takeaway: The conversation doesn't end when the context window fills. That's a bigger deal than another benchmark point.

Claude Moves Into Your Spreadsheets and Browser

Claude for Chrome is now available to all Max users. Claude for Excel is available to Max, Team and Enterprise users with support for pivot tables, charts and file uploads.

➡️ The move: Claude for Excel brings the chatbot to a sidebar inside of Microsoft's app. Claude can navigate complex models, explain formulas, update assumptions across an entire model while preserving dependencies, and trace errors to their source in seconds.

Accuracy on internal evaluations improved 20%, efficiency rose 15%, and complex tasks that once seemed out of reach became achievable.

⚡Why it matters: Most AI assistants live in a browser tab you switch to. Anthropic is embedding Claude directly into the tools where knowledge work actually happens. Through Claude for Chrome, Opus 4.5 gains structured access to the browser's Document Object Model. It can read documentation pages, navigate multi-step flows, fill forms, and "glue together" disparate web interfaces.

🎯 Your Takeaway: The browser extension and Excel sidebar aren't features. They're a distribution strategy.

🔧 TOOL SPOTLIGHT: The Effort Parameter

Claude Opus 4.5 is the only model that supports the effort parameter, allowing you to control how many tokens Claude uses when responding.

What it does: With the new effort parameter on the Claude API, you can decide to minimize time and spend or maximize capability. Set to a medium effort level, Opus 4.5 matches Sonnet 4.5's best score on SWE-bench Verified, but uses 76% fewer output tokens. At its highest effort level, Opus 4.5 exceeds Sonnet 4.5 performance by 4.3 percentage points while using 48% fewer tokens.

How to use it: Use high effort when you need Claude's best work: complex reasoning, nuanced analysis, difficult coding problems. Use medium effort as a balanced option when you want solid performance without the full token expenditure. Use low effort for high-volume automation where speed matters more than depth.

Why it matters for you: If Sonnet 4.5 takes three turns to fix a bug because it keeps missing a subtle dependency, you've burned tokens and, more importantly, your time. If Opus 4.5 solves it in one turn because of superior reasoning, it's effectively cheaper.

The effort dial lets you match model intensity to task complexity. Use low for quick lookups, high for architecture decisions. That's cost control baked into the API.

👉 TRY THIS TODAY

Turn meeting notes into a formatted Excel tracker in under 5 minutes.

If you have Max, Team, or Enterprise access, open Excel and launch Claude for Excel from the sidebar.

1. Paste your raw meeting notes (action items, owners, deadlines) into a blank sheet
2. Ask Claude: "Create a project tracker from these notes with columns for Task, Owner, Due Date, Status, and Priority"
3. Let Claude structure the data, add formulas for overdue items, and format the sheet
4. Review, adjust, and save

Every explanation includes cell-level citations so you can verify the logic. You keep control while Claude handles the tedious formatting work.

✨ THE WIRE

Microsoft and Nvidia invest in Anthropic. The investments boosted Anthropic's valuation to about $350 billion. CNBC

Claude Code hits the desktop app. Claude Code is now available in the desktop app, and with Opus 4.5, it can build more precise plans and execute them more thoroughly. MacRumors

Opus-specific caps removed. For Max and Team Premium members, overall usage limits have increased. Users get approximately the same number of Opus tokens as they had with Sonnet previously. MacRumors

LINKS & RESOURCES

Anthropic's official Opus 4.5 announcement – Full technical details and benchmark comparisons

Claude for Excel documentation – Setup guide and best practices

Simon Willison's first look – Independent developer perspective on real-world performance

VentureBeat deep dive – Business implications and enterprise adoption

Neural Notes
AI that amplifies your value, not replaces it.