Claude Sonnet 5 Just Dropped — Is It Actually Worth It?
No-hype review / est. runtime ~14 min
Cold Open — Hook
[ camera up · no intro · talk straight to lens ]
All right, so Anthropic just dropped Claude Sonnet 5, and the AI internet is doing its usual thing — everybody screenshotting benchmarks, everybody yelling "this changes everything." And look, Sonnet's the one nobody live-tweets, right? It's not the flashy flagship. Opus gets the headlines, Sonnet gets ignored. But here's the contrarian take I keep coming back to: for most of us — the people actually using this thing to draft emails and build little tools — the flagship was always overkill. Sonnet 5 is reportedly hitting numbers two points off Opus 4.8 while costing a fraction of the price. A fraction. So the real question isn't "is it the smartest model." It's "is it the smartest model you'll actually pay for." So in this video, I'm going to put Sonnet 5 through the exact same real-world tests, side-by-side with Opus, and tell you straight up whether it's worth switching. No hype. Let's get into it.
Seven beats, paced the way the channel actually flows — quick what-changed, what's in it for you, two live head-to-head builds, a rant, the honest limits, then a straight verdict.
-
What Actually Shipped (60 Seconds, No Fluff)
Speed-run the real changes — context window, latency, the one headline benchmark number — and skip every line of Anthropic's marketing copy.
-
The Price Tag Nobody's Talking About
Put Sonnet 5's cost per million tokens right next to Opus 4.8 on screen and translate it into what a normal month of actual usage costs you.
-
Test 1 — The Boring Office Stuff That Pays the Bills
Same prompt to both models on an email reply, a messy doc summary, and a spreadsheet cleanup — the work mortals do all day, not a coding demo.
-
Test 2 — Okay, But Can It Build?
Hand Sonnet 5 and Opus the identical one-shot prompt for a little landing page and tool, deploy both live, and compare the output side by side.
-
Simmons Rant — Stop Defaulting to the Flagship
Quick aside on why people reflexively burn money on the biggest model out of habit, and a simple rule of thumb for picking the right one per task.
-
Where Sonnet 5 Falls Apart
The honest part — the exact tasks where it visibly loses to Opus, so you know the ceiling before you go all in and switch everything over.
-
The Verdict — Should You Actually Switch?
Straight answer on who Sonnet 5 is for, who should stay on Opus, and the two-minute setup to swap it in — then subscribe so you catch the next test.
End scene · see you in the next one