Banja
About
Services
Products
Case Studies
Lab
Contact Us
Let us pitch to you

LET'S BUILD
THE FUTURE.

Start a Project
or
Meet Jett
banja.au

We build digital products for people who move fast.

Explore

•About•Case Studies•Blog•Careers•Contact

Services

•Product Design & Build•AI Agents & Automation•Website & Brand Setup

Products

•Boosta

Contact

helloremovethis@andthisbanja.au
50 Miller St
North Sydney NSW 2060

© 2026 Banja Labs. All rights reserved.

Privacy PolicyTerms of Use
Banja Lab

Measured in the open.

The things we build to answer our own questions: experiments, research, and tools, shared in the open. We start with model benchmarks, run on the same work we ship for clients - currently 30 models across 87 tasks, topped by grok-composer-2.5-fast at 85.0% composite.

85.0%
Composite
CodeUISVGWebLawAcct
Benchmarks

Model benchmarks

How frontier AI models perform on the real build and reasoning work we do: code, UI, full websites, marketing pages, dashboards, animations, SVG, and Australian legal and accounting. Measured, with every output to inspect and compare.

Models
30
Tasks each
87
Scoring
Execution + vision
View benchmarks
Ingrid Kessler
Yuki Tanabe
Mara Vieira
Elias Byrne
Noor El-Amin
Designers

Designer benchmarks

Five fictional principal designers, each a fixed and distinct design identity, building the same client briefs across every benchmarked model. See how one point of view holds as the model behind it changes.

Designers
7
Models each
4
Scoring
Cross-family vision
View designers

More experiments land here as we build them.