Base44 adds Claude Opus 5 the day it launches
By Steven Van ·
The new flagship roughly doubles Opus 4.8's Frontier-Bench score while costing half as much per token.
Base44 added Claude Opus 5 the day Anthropic shipped it. It's a one-line announcement, but for a platform that turns a prompt into a full-stack web app, the model doing the building is the entire product, and Opus 5 is a meaningful upgrade over what came before.
Why a better model matters for Base44
Base44 builds complete apps, UI, data, auth, deployment, from a description. That's a demanding, multi-step generation task, and it's exactly what Opus 5 was designed for. Anthropic's new flagship roughly doubled Opus 4.8's score on Frontier-Bench (43.3% vs 18.7%), the benchmark that measures agentic, multi-file engineering rather than isolated snippets. The harder and more interconnected the thing you're asking Base44 to build, the more that jump in capability should show up in the result.
Frontier intelligence at half the price
Opus 5's pitch is roughly Claude Fable 5-level intelligence at half the cost ($5/$25 per million tokens), plus a low/medium/high effort toggle. For a builder that runs the model on every generation, cheaper frontier output means users can iterate more for the same spend, and complex builds become more affordable to attempt. It's the kind of upgrade that quietly improves both quality and unit economics at once.
Part of a busy stretch
Opus 5 lands amid a fast run of Base44 releases, its Components Library, a Superagent Chrome extension, and Wiz security integration, all recently. A stronger base model raises the ceiling on all of it: better first-pass apps that then benefit from the components, security, and agent tooling Base44 has been layering on.
Try it
Claude Opus 5 is available on Base44 now. Describe an app with real moving parts, multiple pages, data that connects, some logic, and see how much closer the first build is to something you'd actually ship.
Sources: Base44 on X, Anthropic.