Anthropic just released Claude Sonnet 5, and it is now the default model for Claude users at no extra cost. The biggest gains are in reasoning, professional knowledge work and the agentic tasks this series runs on. If you already use Claude, you are already on it.
Anthropic calls Sonnet 5 a substantial improvement over the previous Sonnet on reasoning, tool use, coding and knowledge work. On the agentic benchmarks it published, the new model is a strict improvement over the old one.
It also reports lower rates of hallucination and sycophancy than the previous Sonnet. In plain terms, it makes up fewer facts and is less likely to tell you what you want to hear. That is the upgrade that matters most when you are acting on the output.
Four things moved, and they are the four that matter for real work: reasoning, professional knowledge work, agentic workflows, and reliability.
Reliability is the one to notice. Claude is better at checking its own numbers before it answers. That is exactly what you want when a figure is going into a deal.
Sonnet 5 is the new default model. You do not install it and you do not toggle it. Your next new chat is already running on it, so the improvements reach you without you changing a thing.
This upgrade matters because the tasks you run every week get sharper, with no change to how you work. The benchmark is only the proof of it.
These are the property jobs where stronger reasoning and better reliability show up first.
Analysing a deal. Reviewing a rent roll. Building or stress-testing a financial model. Producing a portfolio report. Automating the repetitive workflows that run across a property business.
Every one of those is professional knowledge work with numbers attached, which is the category Anthropic says improved most.
A benchmark is a number on a chart. The upgrade you actually feel is the deal in front of you getting a sharper answer.
When you underwrite, a confident wrong number is worse than no number. A model that checks its own work before answering is the difference between a figure you can act on and one you have to re-check by hand.
There is no new tier and no add-on. If you already pay for Claude, Sonnet 5 is included as the default model, so the better answers cost you nothing extra.
Sonnet 5 is the new default for Claude users. If you already have a plan, it already includes this.
Nothing to install and nothing to toggle. New chats use the new model on their own.
It is already live as the default, so your next new chat is running on it.
Most upgrades this useful arrive as a more expensive tier. This one arrived as the default, at the price you already pay.
These cover the three jobs from the reel: underwriting a deal, reading a portfolio, and turning a repeated task into a workflow. Each one tells Claude to show its working and flag what it is unsure of, so you are using the reliability gain instead of just trusting the output.
Fill in the brackets and paste into a new chat. Start a Project first if it is something you will come back to.
Act as a sceptical property underwriter. I am assessing this deal: - Property: [type, location, size] - Price: [asking price] - Expected rent: [monthly or annual, gross] - Costs I know of: [service charge, agent fees, financing terms, anything else] Do a first-pass underwrite. Return: 1. Gross and net yield, and the cash-on-cash return if I finance it. 2. Every assumption you had to make to get there, listed separately. 3. The two or three numbers that would change the decision most if I have them wrong. Do not invent figures. Where you are missing an input, say so and tell me what to find. Check your own maths before you answer, and flag anything you are unsure about.
Act as a portfolio analyst. Here are the properties I hold: [One row per property: location, purchase price, current value estimate, annual rent, annual costs, outstanding loan and rate.] Analyse the portfolio. Return: 1. Net yield and rough total return for each property, ranked best to worst. 2. Which asset is dragging the portfolio, and why. 3. One question I should be asking about each of the bottom two. Show the calculation behind each number so I can check it. If a row is missing data, list what is missing rather than guessing.
Act as an operations partner for a property business. A task I repeat often is: [Describe the task in a sentence, e.g. reviewing a new listing, chasing a late rent payment, preparing a monthly update for a co-investor.] Design a repeatable workflow for it. Return: 1. The exact steps, in order, that I should run each time. 2. A single prompt I can reuse for the thinking part, with brackets for the inputs that change. 3. Where a human still needs to check before anything goes out. Keep it simple enough that I can run it in minutes. Ask me one or two questions first if the task is unclear.
Save the prompt you use most into a Project's instructions, so every new chat in that Project already knows how you want it to work. That is how a one-off prompt becomes part of your setup.
It is tempting to treat every launch as the thing that finally makes AI useful for property. The model is the engine, and a better engine is welcome.
The advantage that lasts comes from the profile, the Projects and the workflows you build around the model. A stronger model just gives that setup more to work with.
Think of the model as the engine. Sonnet 5 is a better engine, and it arrived without you asking for it.
The driver is still your setup: what you have told Claude about you, and the Projects and workflows you point it at. A better engine only helps as much as the driver behind it.
A launch raises the floor for everyone. What you build on top of it is what pulls you ahead.
Message me directly if you want to talk through how to set up your Projects and connectors for the work you actually do. If property is part of your plan, I can also walk you through the opportunities I am tracking right now.
Message me on WhatsAppSonnet 5 is already working in every new chat you start. The upgrade you actually feel is the one you build on top of it.
A stronger model raises the floor for everyone at once. The reason two investors get very different answers from it is the context each one has set up around it.
That is what this series builds, one prompt at a time, so every upgrade Anthropic ships lands on a business you have already taught Claude to understand.
Every prompt in this 30-day series is one capability. FourthspaceOS bundles all of them into a single product: underwriting, comps, market research, deal sourcing, portfolio tracking, and investor reporting. The product runs natively on the same Anthropic agents and plugins you are learning to use this month.
Waitlist members get founding pricing, early access, and one new prompt delivered each day for the next 30 days.
Founding pricing locks in for waitlist members.
One email a day for 30 days. No spam. Unsubscribe whenever.
The rundown you just read is Day 23 of 30. The full series, daily builds, and live property workflows are on Instagram. Follow for the rest of the run.
If you found this page through the reel, the feed is where the next one lands. One prompt, one build, one workflow at a time, all built for property investors.
@fourthspaceOS · new prompt every weekday
Follow on Instagram ↗Free. Daily. Unfollow whenever.